Professional Documents
Culture Documents
9-7-1999
This Article is brought to you for free and open access by the Harvard Law School at NELLCO Legal Scholarship Repository. It has been accepted for
inclusion in Harvard Law School John M. Olin Center for Law, Economics and Business Discussion Paper Series by an authorized administrator of
NELLCO Legal Scholarship Repository. For more information, please contact tracy.thompson@nellco.org.
ISSN 1045-6333
*John M. Olin Fellow in Law and Economics and SJD Candidate, Harvard Law School.
To
overcome this time inconsistency, over the years durapolists have developed various
business strategies. I have studied the main strategies, their anticompetitive effects and the
ways they were treated by the courts, lawyers and economists. In particular, the various
practices of commitments to future prices, planned obsolescence, tying arrangements, and
leasing are examined.
John M. Olin Fellow in Law and Economics and SJD Candidate, Harvard Law School.
-1-
Introduction.................................................................................................................................................3
II.
II.2.
II.3.
II.4.
III.
III.1.
III.2.
Reputation............................................................................................................................... 22
III.3.
III.4.
III.4.1.
III.4.2.
III.5.
IV.
III.5.1.
III.5.2.
III.5.3.
IV.1.
IV.2.
IV.3.
IV.3.1.
IV.3.2.
Harvard Law School. This paper is based on my thesis paper at Harvard Law School. I am deeply grateful
to my supervisor, Louis Kaplow, for comments and suggestions throughout the writing of this paper. I
have also received helpful comments from F.M. Scherer, Sam Breen, Ciaran McIntyre, Liran Einav, and
Nitsani Chorev. I would also like to thank to David Tadmor for introducing me to the durapolist.
-2-
IV.4.
IV.4.1.
IV.4.2.
IV.4.3.
IV.5.
IV.6.
V.
V.2.
V.2.1.
Framework of Discussion.............................................................................................. 68
V.2.2.
V.2.3.
V.3.
VI.
VI.1.
VI.2.
VI.3.
VI.4.
A.
Introduction
The durapolist1 is an ancient creature in our world and a distinguished member of
the monopolists family. The durapolist is the durable-good monopolist, a frequent guest
in the courtroom, and a favorite subject of study among lawyers and economists following
a seven-page note by Professor Ronald Coase.2 Several scholars have argued that the
durapolist has a limited market power and, therefore, should not be a concern of antitrust
1
2
-3-
agencies.3 Others have indicated that the durapolist engages in certain practices that assist
him in acquiring and maintaining market power. This paper explores the durapolists
practices and how have they been addressed by the courts, lawyers, and economists.
A durable good is a long-lasting good which can be consumed repeatedly. The
monopoly over durable-goods markets is durapolists invaluable asset. In other markets
there are consumable goods (or consumables) that cannot be used more than once, even if
they are characterized by longevity, such as canned food and cigarettes are. The markets
for consumables may be controlled by the durapolists close relatives, the consumablegood monopolists.
durables is functionally different than a monopoly over consumables and these differences
are discussed below.
The unique characteristic of the durapolists is that the durability of their goods
causes them to be inconsistent over time and as a consequence they cannot charge the
static monopoly price. This is the Coase conjecture.4 Historically, however, it is well
known that many durapolists had a tremendous market power and, thus, the Coase
conjecture presumably poses a puzzle: how do durapolists succeed in overcoming their
time-inconsistency problem?
Overcoming the durapolist problem is the topic of this paper and specifically the
practices and methods that durapolists employ in order to do so. My topic might be
3
4
-4-
Simply put,
condemning shady practices, which are employed to overcome the durapolist problem,
may sometimes oppress honestly industrial expressions of superior skill, foresight and
industry, contrary to the goals of the Sherman Act.6
One more preliminary comment about the nature of the durapolists I have
examined should be made. In many respects the differences between a monopoly and a
cartel are insignificant. Both presumably constitute one unit that dictates the quantities or
prices of goods on the markets. Therefore, I considered cartels of coordinated durable-
5
6
15 U.S.C.A. 2.
United States v. Aluminum Co. of America, 148 F.2d 416, 430-431 (2nd Cir. 1945).
-5-
applicability of tying and leasing practices as devices to overcome the durapolist problem.
The uses of these practices by durapolists have gained much attention by the courts,
lawyers, and economists and I examine their treatment. Section VII contains concluding
remarks.
B.
B.1.
-6-
durapolist has supposedly earned the same revenue he would have for selling one-year
bulbs, but has saved the costs of high-volume transactions.7 The fallacy of this intuition is
discussed in this section.
The light-bulb durapolist, as the rest of the durapolists, is perceived to have a timeinconsistency problem and because of this perception he earns less when selling three-year
bulbs than he does when selling one-year bulbs.
production line of durables, the durapolist calculates the optimal output, which will
maximize his profits, and plans his steps in order to attain the monopoly price. This plan
is the durapolists optimal plan. Nevertheless, the durapolist has incentives to abandon the
optimal plan when making subsequent production decisions. In durable-goods markets,
some of the potential demand in period t is satisfied because of past purchases. Therefore,
the demand the durapolist meets in period t is more elastic than the demand he met in
period t-1 and in previous periods. Consequently, the durapolist has incentives to lower
his prices in period t.
The consumer anticipates this behavior, and believes that the optimal plan of the
present moment is generally one which will not be obeyed, and the durapolists future
behavior will be inconsistent with his optimal plan.8
t { 1,2 } the consumer believes the price of the three-year bulb will fall next year, and
-7-
bulbs, the consumer will again have to purchase a bulb next year. Therefore, she does not
have expectations for price-cutting and she is likely to pay the monopoly price.9 It seems,
therefore, that durability is not necessarily what durapolists like.
B.2.
The Model
1
1+ r
, where r[0,] is the discount rate. Namely, 0 1 for any discount rate
between 0 to . Assume also that the secondhand market is functioning perfectly and,
-8-
hence, the costs of consumption are ascribed only to imperfect durability and the costs of
time.
In period t{1,2} the durapolist sells quantity qt of his durables and this quantity is
also consumed in period t. Let the marginal value of consumption for a consumer in
period t be vt = 1 Qt , where Qt is the quantity of consumed durables in period t and the
value of consumption of one new unit of these durables is 1.
In period 2, there is a competing supply of used durables that were purchased in
period 1. The value of this supply is q1 and therefore the marginal value of consumption
in period 2 is v2 = 1 q 2 q1 . This is also the price that the consumer is willing to pay
for q2 new durables in period 2, because they have no prospective value for her. That is,
p 2 = 1 q 2 q1 . In contrast, the durables, which were purchased in period 1, have some
value for the consumer in period 2. Therefore, when the consumer purchases durables in
period 1, she is also willing to pay their discounted prospective value (p2).
Accordingly, the price that the consumer is willing to pay in period 1 is p1=1-q1+p2.
Let us begin in period 2, in which the durapolist chooses q2 in order to solve:
max 2 = ( 1 q 2 q1 )q 2
q2
-9-
1 + 12
, and the price the
2 + 2
( 1 + 12 )2
4 + 22
durapolists profits there. The price the consumer is willing to pay in this period is:
p 2 = q 2 = 21 ( 1 q1 ) = 12 12
1 + 12
2 + 2
-10-
1 + 12
2 = 12
4 + 22
2 + 21 2
=
2
4 + 2
2
This enables us to sum up the durapolists profits for periods 1 and 2, in order to
realize his total profits in terms of period 1:
12
2 + 21 2
( 1 + 21 ) 2
=
+
2
4 + 22
4 + 2
B.3.
This result of the durapolists profits as a function of the durability and the
discount rate can be portrayed in a three-dimensional graph as follows.
In
durapolists
this
graph,
chart,
axis
the
X
3URILWV
'XUDELOLW\
0
Discount
Rate
-11-
light-bulb durapolist presumably failed to see at time zero:11 the durapolists profits
decrease as the durability of his goods increases.
This durability/profit ratio is not linear. The marginal profit of reduced durability
is reducing. Thus, for instance, under a zero discount rate, a reduction of 0.2 durability
from 0.2 to 0 increases the profits by 0.8% (from 0.496 to 0.5), while the same reduction
from 1.0 to 0.8 increases the profits by 3.65% (from 0.438 to 0.454). Simply put, the
incentives to reduce durability are greater when the durability is higher. Or, alternatively,
the incentives of the durapolist to reduce durability are higher than those incentives of his
relative, the consumable-goods monopolist.
Furthermore, the durapolists chart illustrates the impacts of the discount rate on
the demand and supply for durability. A high discount rate means a low present value of
future utility for consumers and producers, i.e., future consumption and future profits are
worth less today. From the consumers point of view, it means that under a high discount
rate she is less sensitive to shifts in durability and more sensitive to shifts in prices.12
11
The following table provides the durapolists profits under various combinations of durability and discount
rate (r):
r
0
25%
66.7%
150%
400%
12
0.2
0.4
0.6
0.8
1.0
0.5
0.45
0.4
0.35
0.3
0.25
0.496
0.447
0.398
0.349
0.299
0.25
0.485
0.438
0.392
0.345
0.298
0.25
0.47
0.427
0.384
0.34
0.295
0.25
0.454
0.414
0.374
0.333
0.292
0.25
0.438
0.4
0.363
0.325
0.288
0.25
The relationship between demand for durability and discount rate is used for different indices which
measure the orders and/or shipments of durable goods as indicators for consumer confidence in the
economy. Examples of such indices are the Commerce Departments monthly reports on orders of durable
goods and the Conference Boards Index of Consumer Confidence. See: Kinsey & Collins [1994].
-12-
Thus, for instance, under a high discount rate, the consumer is willing to pay for a threeyear bulb almost the same amount she is willing to pay for a one-year bulb.
From the durapolists perspective, the discount rate has two main implications.
First, when the discount rate is high future incomes count less and, therefore, the
durapolists incentives to reduce durability are lower than when the discount rate is low.
Under infinite discount rate, durability has no impact on the durapolists profits (0.25 for
any durability). In contrast, under standard economic conditions, i.e., a low discount rate,
the durapolist earns from reducing the durability.
durability are stronger when the durapolists discount rate is lower than the consumers
discount rate. This scenario is plausible when the durapolist is a firm and the consumer is
an individual. It is likely that future profits are valued more for a firm than future
consumption valued for an individual. Under these circumstances, reducing durability
serves also as a device to slow down the decline of the durapolists business that is caused
by the high discount rate of the consumer.13 To illustrate, assume that the consumer has
an infinite discount rate and the durapolists discount rate is significantly lower. The
consumer is willing to pay for the three-year bulb no more than she is willing to pay for
the one-year bulb. As a result, the durapolists profits decline with durability because his
prospective revenue is lower since the bulbs last longer. Therefore, by reducing durability
the consumer is not worse off but the durapolist is better off.
13
Barro [1972] proved this argument in a simple model and argued that the incentives to reduce durability
exist only when the discount rate of the durapolist is lower than the one of the consumer.
-13-
The intermediate conclusion of this discussion so far, is that the durapolist has
incentives to reduce the durability of his goods.
negatively correlated to the discount rate and with to the difference between the discount
rate of the consumer and the durapolist. Simply put, the durapolists chart illustrates the
Coase conjecture and the fundamental belief that by reducing durability the durapolist may
overcome his problem. Less clear, however, is what this durability () is, that by
reducing it the durapolist earns more. The simple model above permits two elementary
interpretations: the value of a durable in the subsequent period and the fraction of durables
that survive consumption and can be used in the next period.
represents the loss of value which occurs in consumption over one period. The following
sections interpret more deeply the notion of durability in this context while observing
durapolists in action. Nevertheless, we will not be fully equipped to do so, unless some of
the assumptions above are highlighted.
First, the durables were assumed to be homogeneous over time. Namely, the
durables produced in one periods are identical to those produced in earlier periods.14
Second, the durability was presumed to be determined before the time of purchase and
could not be changed by the durapolist or the consumer afterwards.15 Third, it was
assumed that the durapolist incurs no production costs, and in particular his production
costs do not increase with durability.16 Fourth, the consumer was assumed to be fully
14
-14-
informed of the durability.17 Fifth, it was also premised that the consumer incurs no
maintenance costs. Sixth, the sales were characterized as discrete, rather than continuous;
namely, the durapolist sells only at the beginning of each period of time, and does not
sell his durables continuously over time.18 Seventh, the secondhand market was premised
to be functioning perfectly and, therefore, durables lose value over time only because of
their imperfect durability and the discount rate.19 Finally, the durapolist was presumed to
have no ability to commit to future prices.20
B.4.
The preceding discussion focused on the durapolists profits and their correlation
to durability and the discount rate. We observed that there is such a correlation, but we
have not examined the effects, if any, on social welfare. This subsection analyzes the
impacts of durability and discount rate on social welfare.
We begin by finding the social surplus that results from transactions between the
durapolist and the consumer. In period t the consumer purchases qt goods and consumes
Qt goods, where Qt includes qt and goods from the previous periods that survived their
consumption.
17
-15-
-16-
S 2 = 41 q1 81 2 q12 + 83 .
Accordingly, the aggregate social surplus in terms of period 1 is:
S = S 1 + S 2 =
= q1 ( 1 + 41 ) q12 ( 21 + 81 2 ) + 83
1 + 12
and, therefore we can present
2 + 2
( 2 + 8 ) + 83 .
S =
(
1
)
+
4
2
2
2
2
+
+
clearly shows
the
connection between
durability, discount
'XUDELOLW\
-17-
'LVFRXQW
5DWH
6RFLDO 6XUSOXV
21
The following table provides the aggregate social surplus under various combinations of durability and
discount rate (r):
r
0
25%
66.7%
150%
400%
0
0.75
0.675
0.6
0.525
0.45
0.375
0.2
0.4
0.6
0.8
0.794
0.710
0.626
0.543
0.456
0.375
0.826
0.735
0.645
0.555
0.465
0.375
0.843
0.75
0.656
0.562
0.469
0.375
0.848
0.754
0.659
0.565
0.47
0.375
-18-
1.0
0.8438
0.75
0.6563
0.5625
0.4688
0.375
enables consumers to buy more of his goods and, therefore, their surplus increases, while
the durapolist surplus is negligible because it represents merely a transfer from the
consumers.
consumables and this is why generally the social surplus increases with durability.
Nonetheless, durability also causes the durapolist to limit his output in period 1 because he
factors the impact of this output on his sales in period 2. As a result, durability also has a
negative impact on social surplus and this is why we see a negative correlation between
durability and social surplus at high levels of durability.
The numbers behind the social surplus may shed more light on this outcome.
Under a discount rate between zero and 25%, reductions of durability from a level of 0.8
to a level of zero render a loss of 12-13% in the social surplus and from a level of 0.8 to a
level of 0.6, a loss of approximately 0.5%. In comparison, an increase of durability from
0.8 to a perfect durability (1), under these levels of discount rate, leads to a loss of social
welfare of approximately 0.5%. This implies that higher durability is not always socially
beneficial and the subsequent discussion develops this point further.
To the extent that distributive considerations are relevant, higher durability shrinks
the portion of the surplus that the durapolist can grasp because of his time inconsistency.22
As mentioned, the portion of social surplus, which the durapolist can capture, may impact
22
The following table provides the division of the social surplus in the critical values of durability (0, 0.8, 1)
when the discount rate is zero:
Durability ()
0
0.8
1
Social Surplus
0.75
0.8483
0.8438
Durapolist Surplus
0.5 (66.7%)
0.454 (53.5%)
0.4375 (51.8%)
-19-
Consumer Surplus
0.25 (33.3%)
0.394 (46.5%)
0.406 (48.2%)
his incentives to invest in entry to markets or to create new markets and this can be
another social cost of durability.23
C.
C.1.
A Matter of Confidence
In the famous Alcoa case, Judge Learned Hand commented that [limitations to
market power] also exist when a single producer occupies the whole market: even then, his
hold will depend upon his moderation in exerting his immediate power.24 Judge Hand
premised that the durapolists business plans are transparent to the consumers and, hence,
his market power depends on his conduct rather than on the consumers perception of this
conduct. The essence of the durapolist problem, however, is his low credibility for being a
consistent creature and the expectations that he will lower his prices. If credibility is the
problem, logic dictates that the durapolist should find ways to commit to future prices.
This logic follows a fundamental economic rule: a committed player can do at least as well
as without a commitment.25
durapolists have developed in order to gain credibility and the public confidence.
23
24
25
Lee & Lee [1998] provided an analytical framework of this point and particularly, the potential of
underinvestment in R&D activities in durable-goods markets. See, also Waldman [1996].
United States v. Aluminum Co. of America, 148 F.2d 416, 425-426 (2nd Cir. 1945). For an economic
analysis of the Alcoa case, see: Gaskins [1974]; Swan [1980].
Needless to say, a rational player would not commit to unprofitable conduct, or conduct that would make
him worse off. See, generally: Schelling [1960]; Dixit & Nalebuff [1991], at 142-167. For a formal
analysis of commitments by durapolists, see: Hart & Tirole [1988].
-20-
26
-21-
C.2.
Reputation
Had the durapolist had a reputation for keeping promises, he would have had no
obstacles to his market power. Unfortunately for the durapolist this is not the case. In
general, his reputation is for breaking promises in order to maximize his short-run profits.
This is his time-inconsistency problem. The durapolist is unlikely to sacrifice short-run
profits even for greater long-run profits. One way to explain this irrational behavior is
through the repeated decisions the durapolist must make regarding his pricing policies.
Essentially, every time the durapolist sells his durables he has to decide whether to follow
his optimal plan or to deviate from it and to charge the monopoly price according to the
27
In expectancy there are 50 high-value consumers and, therefore, when the price is $5 the durapolists
profits are expected to be $250.
It should be noted that this example assumes a continuous supply of bulbs, in contrast to the model
presented in section B. If the supply is discrete, under a policy of non-commitment the durapolists
profits will be between 200 and 250
28
Furthermore, tracks of past purchases enable the durapolist to price discriminate and, therefore, high-value
consumers will be unlikely to reveal themselves by paying the high price. Fudenberg & Tirole [1998]; For
further implications, see also chapters F-E. A less persuasive argument concerning the high-value
consumers is that they are unlikely to pay the monopolistic price because they understand that only this
willingness to pay renders the high level of prices. See, for example Posner [1976], at 203. This argument
fails to recognize the consumers collective-action problem.
-22-
residual demand.
optimal plan is costly and therefore, may seem undesirable to the durapolist.29 Some
durapolists, however, have seen beyond the immediate horizons and thus, gained much
prosperity. A prominent example of a durapolist who presumably maintains market power
through reputation is the international diamond cartel, led by DeBeers.30
A Diamond is Forever, so reads the DeBeers slogan, which leads the ultimate
durapolist and manages its reputable time-consistency. Cecil Rohdes, the founder of
DeBeers, devised the strategy of the diamond cartel in the 1880s, and his successor,
Ernest Oppenheimer, developed and perfected it.
simplicity: persistence in restricting the number of diamonds released into the market and
preventing the nominal prices from going down. Thus, for example, several studies of the
industry noted that the quantity of diamonds sold on the market followed the prospective
number of wedding engagements in any given year. 31
29
30
31
An interesting analogy between the durapolist and a predator can be drawn. Both a predator and a
durapolist who sticks to his optimal plan sacrifice short-run profits for the long-run profits. A predator
lowers his prices to punish competitive behavior and thereby, incurs losses. Similarly, a durapolist who
does not lower his prices does not earn as much as he could have. One main difference between the two
is in the number of short-run unprofitable decisions. A predator needs to lower his prices only once (or
very few times) to obtain market discipline, while the durapolist has to make a costly decision every
time he sells his durables. For the repeated game of the predator, see generally: Kreps & Wilson [1982];
Ordover & Saloner [1989].
Again, for the purposes of this paper, durable-goods cartels are considered as durapolists. See subsection
C.5.2 below.
The cartels success is facilitated by a complicated system of stockpiles, production quotas, tight price
monitoring and sophisticated advertising. Much of this system is regulated through the cartels Central
Selling Organization (CSO), by which it determines who can buy which stones and how much each buyer
must pay. See: Spar [1994], at 39-87; Lenzen [1970]; Anonymous, How De Beers dominates the
diamonds, The Economist, Feb. 23, 1980, at 101; THE ECONOMIST INTELLIGENCE UNIT, DIAMONDS 1988,
SPECIAL REP. NO. 1126 (DEC. 1987); THE ECONOMIST INTELLIGENCE UNIT, DIAMONDS: A CARTEL AND
ITS FUTURE, SPECIAL REP. NO. M702 (AUG. 1992).
-23-
Indeed, in several instances, adhering to this strategy was very costly in the short
run,32 but DeBeers affluence proves that a farsighted durapolist eventually gains more. In
short, the diamond cartel has succeeded in overcoming the durapolist problem, by resisting
the temptation of increasing output and, thereby, building reputation. In this, DeBeers has
succeeded where others have failed.33
C.3.
Diamonds are forever, but most durables are not. As a consequence, a gardenvariety durapolist does not have to precommit that his prices will never go down and a
weaker commitment suffices.34
32
33
34
In 1981, for instance, the diamond sales fell 46% below their level in 1980, and left DeBeers with a
stockpile estimated to be equal to a normal years worth of sales. In the process, DeBeers spent between
$700 million and $1 billion of its own cash reserves to support the diamond prices. Spar [1994], at 56.
Likewise, in addition to the regular premium DeBeers was paying for Soviet/Russian stones to prevent
them from entering the world market, in 1990, it gave the Russian government a $1 billion hard-currency
loan for exclusive rights on selling the Russian diamonds. Spar [1994], at 64-73, 78-87.
Again, in this paper I generally do not distinguish between a monopolist and a cartel. See subsection
C.5.2.
In Allen-Myland, the Court of Appeals for the Third Circuit intuitively linked the durapolists credibility,
which was assumed in Alcoa, to durability:
Alcoas analysis is persuasive Refined aluminum can be melted down and reused
repeatedly, and in any event, products made with it may last for decades before they are
scrapped and the aluminum is recycled. It therefore may have been quite difficult for Alcoa to
estimate future supply and demand for aluminum ingot over a long period of time with
sufficient accuracy to maximize its profits by manipulating the supply of virgin ingot it
produced
Computers, however, have considerably more limited lives than aluminum ingot. Technology
and price/performance ratios have been advancing so rapidly in the computer industry that
used machines cannot be re-leased indefinitely. Accordingly, a powerful manufacturer like
IBM was in a position to maximize its profits by carefully controlling the number of
mainframes that would later appear on the used leasing market. Allen-Myland v. International
Business Machines Corp., 33 F.3d 194, 203 (3rd Cir. 1994) (footnote omitted), cert. denied,
513 U.S. 1066 (1994).
-24-
Conceptually, it is a wonder why the consumer who purchases a one-year bulb has
more confidence in the price than when purchasing a three-year bulb. After all, if the
light-bulb durapolist can lower his prices every year, he can also lower them every week,
day, or in the twinkling of an eye.35 There are various explanations for the consumers
confidence in short periods of time.36 One straightforward explanation is that in reality
price adjustments take time. Firms incur costs in deciding on price changes, generating
new price lists, and catalogs, changing price tags, etc.37
characterized by short-run rigidity. Under such circumstances, when the durables lifetime
is relatively short, a consumer can anticipate that she will save little, if at all, by
postponing the purchase of a short-life good. This is to say, it is easier for the durapolist
to commit to future prices when the durability of his goods is lower.
35
See supra note 27. In the model presented in section II the consumer has such confidence because it was
assumed that the selling was discrete. Namely, the durapolist sold his goods only at the beginning of each
period, e.g., only on January 1st of every year. There are several economists who assert that this
assumption is unrealistic and, therefore, they suggest that reducing durability does not improve the
position of the durapolists. See, for example: Stokey [1981]; Kahn [1986].
36
Denicolo & Garella [1999], for example, provided a general model of rationing in durable-goods markets,
which explains how the durapolist can price-discriminate his consumers over time without a commitment
to future prices:
In a two-period model, [when the goods are not perfectly durable,] buyers who have been
rationed in the first period will carry their demand over to the next one. This shifts the secondperiod demand function upward, reducing the monopolists incentive to cut the second-period
price. Rational customers thus have less incentive to postpone purchase, which improves the
first-period demand. As a result the monopolist can better discriminate between high-value
and low-value customers, and total discounted profit may rise But with an infinite horizon,
when the length of periods shortens to zero, rationing would become ineffective. (at 44.)
37
Another popular explanation for price rigidity in concentrated markets is the kinked demand curve,
which was offered by Paul Sweezy [1939] (and simultaneously by a pair of British economists, Hall &
Hitch), and popularized by Stigler [1947] who cast some doubts by an empirical study which concluded
that [t]he kink is a barrier to changes in prices that will increase profits, and business is a collection of
devices for circumventing barriers to profits. For various discussions about price rigidity, see: Keynes
[1936]; Carlton [1986]; Tirole [1988], at 253-256, 265-268; Baumol, Panzar & Willig [1988], at 12-13,
405-428. See, also Coase [1972], at 147:
What a consumer has to fear is an increase in supply during the period which [s]he (or
someone to whom [s]he transfers the good) is deriving services from the good. The less
-25-
C.4.
durable the good, the shorter is this period. But the shorter the period that the supplier has in
which to increase supply, the greater will be the additional costs of increasing supply.
38
See, generally: Cooper & Fries [1991]; Hovenkamp [1994], at 4.6d:
Buyers may think price protection clauses protect them from subsequent price reductions that
might be given to other firms. If A buys today at a price of $50, and tomorrow the seller sells
to B at a price of $45, A will be entitled to a refund of $5. Nonetheless, such clauses are often
a sign not of hard customer bargaining The clauses effectively make discriminatory price
reductions very expensive.
39
MFC clauses and low price guarantees are different in a setting of more than one seller. An MFC clause
provides the customer with an assurance that the seller will not give to another customer a better price. In
contrast, low price guarantee provides the customer with an assurance that no seller will give a better
price to anyone. When there is one seller in the market, as it is generally assumed here, there is no
-26-
difference between these two practices. For analysis of low price guarantee, see Eldin [1997]. For an
analysis of both practices and their respective legal implications, see: Sargent [1993].
40
See, for example: Cooper [1986]; Tirole [1988], at 330-332; Simons [1989]; Hovenkamp [1994], at 4.6d;
Baker [1996]; Edlin [1997]; Hay [1999].
41
See, for example: United States v. Eli Lilly & Co., 1959 Trade Cas. (CCH) 69,536 (D.N.J. 1959); In the
matter of Ethyl Corporation, 101 FTC 425 (1983); United States v. Vision Serv. Plan, 7 Trade Reg. Rep.
(CCH) 50,775 (D.D.C. 1994); United States v. Delta Dental Plan of Arizona, Inc., 7 Trade Reg. Rep.
(CCH) 50,767 (D. Ariz. 1994).
42
In the matter of Ethyl Corporation, 101 FTC 425 (1983), revd sub nom., E.I. du Pont De Nemours & Co.
v. FTC, 729 F.2d 128 (2d Cir. 1984) (the appeal court vacated the Commissions decision and refused to
hold that 5 could be violated by non-collusive, non-predatory and independent conduct); Ocean State
Physicians Health Plan v. Blue Cross & Blue Shield, 883 F.2d 1101, 1111 (1st Cir. 1989), cert. denied,
494 U.S. 1027 (1990); Blue Cross & Blue Shield United of Wis. V. Marshfield Clinic, 65 F.3d 1406, 1415
(7th Cir. 1995) (Posner, C.J):
This is said to set a floor underneath these physicians prices, since if they cut prices to their
other patients their reimbursement from the Clinic will decline automatically. This is an
ingenious but perverse argument. Most favored nations clauses are standard devices by
which buyers try to bargain for low prices, by getting the seller to agree to treat them as
favorably as any of their other customers It is not price-fixing Perhaps, as the Department
of Justice
believes, these clauses are misused to anticompetitive ends in some cases; but
there is no evidence of that in this case.
43
Aghion & Bolton [1987] (vertical contracts as a barrier to entry); Butz [1990] (a formal analysis of MFC
clauses as a device to overcome the durapolist problem); Png [1991] (the same); OBrien & Shaffer
[1992] (a technical analysis of the profitability of vertical arrangements, including MFC clauses, for a
monopolist); McAfee & Schwartz [1994] (MFC clauses and price rigidity); Goldberg & Greenberg
[1995] (analysis of the Ocean State case).
-27-
more favorable impression on the courts and they generally discard it and resist
acknowledging that it might be used to monopolize markets.44
The courts usually dismiss the allegation regarding the anticompetitive effects of
MFC clauses because they tend to believe that a low-price guarantee benefits the
consumer. As discussed, this belief has no sound foundation.45 The durapolist may use
MFC clauses in order to overcome his time-inconsistency, as competing firms may used
these clauses in order to facilitate coordination. Nevertheless, it is unclear whether, from
the durapolists viewpoint, this practice is as good as other practices, which he can employ
to overcome his problem. Yet, there is no good reason not to infer that the practice
violates antitrust laws, even though its usage by durapolists is rare.
44
The health insurance industry has contributed most of the cases that raised the question of monopolization
by MFC clauses. For surveys of the leading cases, see Celnicker [1991]; Dennis [1995a; 1995b]. See,
also: Kitsap Physicians Serv. v. Washington Dental Serv., 671 F. Supp. 1267, 1269-1270 (W.D. Wash.
1987) (the court opined that an MFC clause makes good business sense and it is procompetitive);
Reazin, at 971 n. 30:
In Ocean State, Blue Cross conceded its monopoly power... The only question was whether
Blue Cross violated section 2 [of the Sherman Act]. By contrast, the most favored nations
clause here is not itself challenged as unlawful monopolization. Rather, it is only considered
as evidence of, or as contributing to, Blue Cross' market or monopoly power. We need not
reach the question addressed in Ocean State of whether use of the most favored nations clause
could itself violate section 2.
45
See, Dixit & Nalebuff [1991], at 102:
The neatest trick is enforcing price collusion through a punishment guarantee [i.e., an MFC
clause,] all in the name of competition.
46
See: Sargent [1993]. Dixit & Nalebuff [1991] illustrated the strategy of MFC clauses with the deceased
Crazy Eddy and its main competitor, Newmark & Lewis. For this purpose, they quoted one of Newmark
& Lewis advertisements:
-28-
overcoming the durapolist problem, or that the durapolist has more profitable practices.
The turbine-generator industry may provide an example for a durapolist who used this
practice.
In 1960, there were three firms in the United States that manufactured electricturbine generators: General Electric (GE), Westinghouse, and Allis-Chalmers. These three
players had controlled the industry since the end of the nineteenth century, when steam
was used to generate electricity. Generators are extremely durable and the three had
maintained their prices through price-fixing agreements.
apart, when the firms were fined and their executives were sent to prison.47
As a
consequence, a price war was spurred, Allis-Chalmers was driven out of the market in
December 1962, and left the pie to be divided between GE and Westinghouse.
This
almost unprofitable situation was stabilized in May 1963, when GE announced a new
pricing policy. This new policy included three elements: (1) the pricing structure was
47
If, after your purchase, you find the same model advertised or available for sale for less
(confirmed printed proof required) by any other local stocking merchant, in this marketing
area, during the lifetime of your purchase, we, Newmark & Lewis, will gladly refund (by
check) 100% of the difference, plus an additional 25% of the difference, or if you prefer,
Newmark & Lewis will give you a 200% gift certificate refund (100% of the difference plus
an additional 100% of the difference, in gift certificates. (at 102-103.)
It is not implied that MFC clauses cannot present high competitiveness, although they are likely to
facilitate tacit collusion in concentrated industries. Nevertheless, it is unclear what can be the purpose of
a general use of this practice by a durapolist, other than commitment to future prices.
GE, Westinghouse and Allis-Chalmers were exposed to a long series of antitrust suits. The aggregate
amount of the claimed damages was estimated to be nine billion dollars (treble damages of three billion
dollars.) Most of the suits were settled out of the courtroom. Interestingly, the settlements contained
MFC clauses, under which the companies promised that the most beneficial out-of-court settlements
would become available to all plaintiffs, regardless of when they settled. Sultan [1974], at 84-124. Ralph
Sultan [1974, 1975] provides the thorough analysis of the industry. The description of the MFC practice
can be found in United States v. General Electric Co., 1977-2 Trade Cas. (CCH) 61,659 (E.D. Pa. 1977).
For documentation of the price-fixing conspiracy see: Herling [1962].
-29-
simplified; (2) discounts were eliminated and a price book was published; and (3) a broad
anti-discrimination clause guaranteed the consumers that for six months from the purchase
time they all will be refunded if GE will sell below its list price. To increase the
consumers trust in its commitment, GE hired the services of a public accounting firm to
audit its policy and, in particular, the equal treatment for all its customers.
Westinghouse was less creative than GE and instead of developing its own policy,
it copied GEs policy and its price book. GE and Westinghouse persisted with this
practice until 1973 and succeeded in maintaining high profits despite the durability of their
products.
The case of GE and Westinghouse does not comment conclusively on the practical
profitability of MFC clauses for durapolists. The duopolistic structure of the industry and
the inelasticity of the demand complicate to some extent the analysis of the practice. Yet,
two issues should be highlighted. Firstly, GE dominated the market48 and, therefore, it
could be viewed as a durapolist.
boundary for GE and Westinghouse, their purpose of stabilizing the duopoly does not
change the fact that they were virtually equivalent to a solution to the durapolist problem.
48
According to the estimate of cumulative kilowatts shipments of the three manufacturers over the years
1898-1963 the market share were as follows: GE 61.4%; Westinghouse 31.2%; Allis-Chalmers
7.4%. Sultan [1974], at 176. After the exit of Allis-Chalmers from the market, the market shares for the
years 1964-1970 were: GE 57.3%; Westinghouse 33.1%; Import 9.1%; Allis-Chalmers 0.5%
(returned to the market in 1970). Sultan [1975], at 228.
-30-
C.5.
Unlike legal doctrines, economic creatures are flexible. Any profitable strategy
has many variants which may be classified by the law as distinctive with different legal
implications although their mechanisms are the same. In this respect, the practice of MFC
clauses is not an exceptional and appears to have many widespread variants. Three of
these variants are analyzed below: (1) returns and buybacks; (2) self-enforcement
mechanisms of cartels; and (3) raising rivals costs.
Returns and buybacks may seem to impose a very severe penalty on the
durapolist which is the high price the durapolist charged before the price reduction.50
This perception is usually wrong because generally the consumer buys only
durables she needs (e.g., car, refrigerator or a stereo) and, therefore, after the refund she
49
For instance, the MFC clause of Newmark & Lewis, at supra note 46, provided a multiplier of 1.25 or 2.
-31-
will again buy the durable, albeit at the lower price. Thus, the durapolists penalty is
approximately the price reduction he set, as the penalty under MFC clauses without the
multiplier.
returns and buybacks represent is relatively impaired because of the transaction costs they
are associated with. Under an MFC clause, a consumer who bought a refrigerator and
finds that the durapolist lowered his prices can almost costlessly claim the difference. In
contrast, under a policy of returns or buybacks, the consumer will lose a workday to
arrange the delivery of her refrigerator, some of her groceries will spoil and eventually she
will have an identical refrigerator, although for a lower price. In short, the strategies of
returns and buybacks as devices to overcome time-inconsistency seem to be dominated by
the strategy of MFC clauses.
Furthermore, unlike MFC clauses, returns and buybacks have important economic
functions other than restricting competition.51 The practice of returns, for example, is a
standard practice in the retailing markets as a form of insurance by manufacturers to the
retailers and in this sense may be perceived as a substitute to consignment transactions.52
Moreover, return policies may constitute another parameter in which firms
compete with each other. While MFC clauses may present tough competition in prices,
50
Coase [1972] suggested that a buyback policy may provide the consumers with a commitment to future
prices (at 145, 148). For an analysis of the potential anticompetitive effect of returns polices, see: Marvel
& Peck [1995]; Padmanabhan & Png [1993; 1997]. See also: Fudenberg & Tirole [1998].
51
See, Eric N. Berg, Building Good Will through Guarantees, N.Y. TIMES, Nov. 23, 1989, at D1; Janice M.
Horowitz, When Gifts Become Problems, N.Y. TIMES, Dec. 24, 1989, at C9. See, also Coase [1972], at
149:
[T]he supplier of a durable good may agee to buy it back at some specified price in the future
because consumers are willing to pay for this reduction in risk.
52
The distribution segment of the book industry, for example, is traditionally based on return policies.
-32-
returns policies also present competition in quality.53 Collusion is less likely when there
are more parameters to coordinate and especially non-quantified ones such as quality.
Returns, therefore, are unlikely to serve as devices to impair competition. Therefore, the
use of returns in these contexts (insurance for retailers and competing sellers) should be
distinguished from the case of a durapolist (one seller) who sells his goods to end users
and provide them with a guarantee of returns.
For instance,
53
54
See: Bob Tedeschi, Online Sales can be Messy, Especially those Pesky Returns, N.Y. TIMES (interactive
ed.), Aug. 23 1999; Berg, supra note 51 (the article was written before the era of the e-commerce):
Nowhere have guarantees become more prevalent than in the mail-order
catalogue business, in which it has become almost standard to permit customers
to return garments and other goods at any time for any reason.
See subsection D.4. Lee & Lee [1998] showed that this internalization may render underinvestment in
innovation. In their model durability may discourage R&D activities because a durapolist who introduces
a new generation of products has to compensate consumers according to their purchase histories. See,
also: Borland, 1992 Annual Report:
The Companys gross margins can be strongly affected in particular in periods by aggressive
pricing strategies and return privileges associated with new product introductions and
upgrades. (cited by Padmanabhan & Png [1997])
This argument is more complicated in network industries because buybacks in these industries may
provide the durapolist with the critical mass that is required to make a present network obsolete. Thus, in
network industries buybacks may facilitate planned obsolescence.
-33-
profitability of this strategy may increase, if the durapolist can employ the used goods to
lower his marginal costs (e.g., using replacement parts) or by increasing the costs of the
players in the secondhand market.55
To sum up this point, generally, returns and buybacks do not seem to operate as
devices to overcome the durapolist problem, i.e., to acquire and maintain monopoly
power. Indeed, returns and buybacks may be used as a commitment. However, the
credibility of these strategies is not always high and, in general, they are dominated by
other strategies. Furthermore, in many cases these practices appear to serve functions
other than a commitment to high prices.
55
-34-
The theoretical significance of this point is that some cartels are better at
maximizing profits than monopolists are and, thus, the social cost of cartels in durablegoods markets may be greater than under the rule of a durapolist.
The practical
significance is less impressive because cartels are illegal anyway and, thus, the practical
implications of this point are limited to the valuation of damages.
56
For discussions about self-enforcement mechanisms of cartels, see: Priest [1977]; Ayres [1987]; Brodley
& Ma [1993].
57
Ausubel & Deneckere [1987] and Gul [1987] provided formal proofs for this argument.
58
See also subsection D.5.
-35-
To illustrate, consider the war of Eastman Kodak against its competitors in the
aftermarket of the photocopiers.59 Kodak tried to drive its competitors out of the market
by extorting agreements from its equipment manufacturers not to sell parts that fitted
Kodaks equipment to anyone other than Kodak. The courts, which examined Kodaks
war against its competitors, found this strategy to be a violation of 2 of the Sherman Act
and, as a consequence, these contracts were invalidated.
An alternative strategy that Kodak could have employed and thereby attract less
attention from the courts is MFC agreements. By signing MFC agreements with the
equipment manufacturers, Kodak could have raised the costs of its competitors in the
aftermarket and eliminate the trading of its used copiers.60
Indeed, the strategy of raising rivals costs was suggested by several economists as
a device to overcome the durapolist problem.61 The technique itself in the general context
of a manufacturer who is interested in MFC clauses in order to erect entry barriers was
59
Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992); Image Technical Services v.
Eastman Kodak Co., 125 F.3d 1195 (9th Cir. 1997). See, also: Mackie-Mason & Metzler [1999].
60
It should be noted that Kodaks main purpose was to acquire exclusivity in the aftermarket, and this was
the focus of the case. Yet, it seems that another purpose was to eliminate the secondhand markets for
Kodaks copiers since it also quadrupled some of its fees to begin servicing a copier bought from
someone other than Kodak. Mackie-Mason & Metzler [1999], at 401.
Klein [1993] suggested that MFC clauses between Kodak and its customers could have limited the
problem of Kodaks market power in the aftermarket. Such clauses would have forestalled the power of
Kodak to raise the maintenance costs for its existing customers without raising these costs for its new
customers. This sort of price discrimination was a main concern of the Kodak courts. One of the
implications of Kleins suggestion is that MFC clauses between a seller and its customers regarding
maintenance services for its primary goods may limit the sellers ability to eliminate the secondhand
market. This argument holds only if the market for the primary goods is competitive. In the Kodak case,
in contrast to the assumption of the Supreme Court, the market was not competitive. Kodak and Xerox
shared a duopoly in the market for high-volume copiers.
61
See infra note 99 and accompanying text. For a discussion about the strategy of raising rivals costs in the
context of MFC clauses see: Krattenmaker & Salop [1986], at 275 n. 207; Baker [1996], at 523-526. For
a technical analysis see: Salop & Scheffman [1987] (implicit analysis); Ordover, Saloner & Salop [1990].
-36-
condemned by several courts,62 but unfortunately other courts did not acknowledge its
anticompetitive effects.63
D.
D.1.
In Grinnell the Supreme Court stated that [t]he offense of monopoly under 2 of
the Sherman Act has two elements: (1) the possession of monopoly power in the relevant
market and (2) the willful acquisition or maintenance of that power as distinguished from
growth or development as a consequence of a superior product, business acumen, or
historic accident.64
durapolists, which involve a business acumen but not necessarily a superior or desirable
product, and which may be intended to acquire and maintain market power.
Technology manipulation, and particularly suppression of technology, is a favorite
theme of urban legends and spooky tales. The accumulation of solar-technology patents
by the oil companies, the introduction of terminator seeds by Monsanto, and the
suppression of a pain-control device by Johnson & Johnson are only a few examples for
62
See, for example: Reazin v. Blue Cross & Blue Shield of Kan., Inc., 899 F.2d 951, 970-971 (10th Cir.
1990), cert. denied, 497 U.S. 1005 (1990) (the jury could reasonably have concluded that [the mostfavored-nations clause contributed to] Blue Cross power over price).
63
Ocean State Physicians, at supra note 42.
64
United States v. Grinnell Corp., 384 U.S. 563, 570-571 (1966) (emphasis added).
-37-
such stories.65 Nevertheless, such schemes are generally not looked upon favorably in the
courtroom, although durability is associated with competition at least implicitly.66
Traditionally, courts have refused to interfere in business decisions concerning the use of
technologies mainly because of the involved intellectual property rights.67
These
implications of the antitrust intellectual property intersection are beyond the scope of
this paper.
The section examines whether there are justifications on antitrust grounds to
interfere in business decisions concerning durability.
65
For the solar-technology patents see: Jensen [1997], at 102-103; Berman & OConnor [1997]. For the
cases of the pain-control device and Monsantos terminator seeds see D.3.2 below.
66
See, for instance: National Society of Professional Engineers v. United States, 435 U.S. 679, 695 (1978):
The assumption that competition is the best method of allocating resources in a free market
recognizes that all elements of a bargain quality, service, safety, and durability and not just
the immediate cost, are favorably affected by the free opportunity to select among alternative
offers.
67
See: Continental Paper Bag Co. v. Eastern Paper Bag Co., 210 U.S. 405 (1908) ([T]he very essence of
the right conferred by the patent, as it is the privilege of any owner of property to use or not to use it,
without question of motive. at 429.) Hartford-Empire Co. v. United States, 323 U.S. 386, 432-33,
clarified, 324 U.S. 570 (1945):
A patent owner is not in the position of a quasi-trustee for the public or under any obligation
to see that the public acquires the free right to use the invention. He has no obligation either to
use it or to grant its use to others. If he discloses the invention in his application so that it will
come into the public domain at the end of the 17-year period of exclusive right he has fulfilled
the only obligation imposed by the statute. This has been settled doctrine since at least 1896
(footnote omitted).
Cf. Special Equipment Co. v. Coe, 324 U.S. 370 (1945) (Justice Douglas, dissenting). See also:
Automatic Radio Mfg. Co. v. Hazeltine Research, Inc., 339 U.S. 834 (1950); SCM Corp. v. Xerox Corp.,
645 F.2d 1195, 1204 (2nd Cir. 1981):
No court has ever held that the antitrust laws require a patent holder to forfeit the
exclusionary power inherent in his patent the instant his patent monopoly affords him
monopoly power over a relevant product market.
-38-
D.2.
68
69
-39-
economic term for intentional reduction of durability, refers, therefore, to both reduced
built-in durability and reduced variable durability.70 These two methods of reduction of
durability are inherently different in their causes and in their impacts and hence require
distinctive discussions.
D.3.
70
71
For discussions regarding reduced built-in durability, see: Bulow [1986]. For reduced variable durability
see: Bartkus [1976] (legal implications of alleged reduced variable durability with a focus on the
computer industry); Rust [1986] (formal analysis of reduced variable durability); Waldman [1993]; Lee &
Lee [1998].
Baker [1995]; A representative line of this skepticism is the Swans independence result, which
presumably rebuts any relationship between market structure and quality or durability of products. See:
Swan [1970a, 1970b, 1971]. For a survey of the literature in this issue until 1982 and the flaws of Swans
independence result, see Schmalensee [1979]; Liebowitz [1982].
-40-
Furthermore,
reduced built-in durability may be coupled with complicated business circumstances and
ambiguous market conditions which make its undesirability very vague. For example, the
costs of higher built-in durability might be higher than what consumers are willing to pay.
In the model presented in section B, it was assumed that the production costs are zero for
any durability.
durability and thus higher durability does not always meet the demand either because of
imperfections of the capital markets or because the rate of the increase in the production
costs is higher than the rate of the increase in durability. Fixed costs, such as investments
in R&D activities, amplify of course these effects.74
72
See for example: Williamson [1975], at 176-233; David [1985]; Klemperer [1987]; Arthur [1994]; Lee &
Lee [1998]; Hendel & Lizzeri [1999].
73
See: Gilbert [1981]; Gilbert & Newbery [1982]; Karp & Perloff [1996]. See also: Kaysen [1956], at 78-91
(description of the fashions United Shoe Machinery used its patents to maintain its market power).
74
When the capital markets are imperfect the consumer who is willing to pay x for a one-year bulb, for
example, will not always have the financial ability to pay 3x for a three-year bulb. For a formal analysis
of a durapolist with increasing costs, see Kahn [1986]. See also Coase [1972], at 149: [A] reduction in
durability may enable a supplier to provide a given stream of services at lower cost.
-41-
Apart from these complications, when the consumer is reluctant to pay more for
higher durability, the existence of good justifications for interference in business
judgement may seem dubious.
As discussed, such reluctance may stem from a high discount rate or from
anticipating inconsistent pricing behavior of the durapolist over time.
In addition,
Information asymmetry and other market imperfections might also render a reluctance in
the consumer to pay more for durability. For example, if the consumer knows the price
but her information concerning the durability is imperfect, she may be unwilling to pay a
higher price for higher durability.75 It should be noted that while time inconsistency and
information asymmetry are reasons which are endemic to the durapolists conduct, a high
discount rate is not related to the durapolist conduct. This suggests that circumstances of
time inconsistency and information asymmetry may justify interference in business
decisions concerning reduction of built-in durability. Reducing built-in durability in order
to overcome the time-inconsistency problem can be regarded as acquiring and maintaining
monopoly power as a consequence of illegitimate business acumen, coupled with an
inferior product.
business acumen and inferior products because there are many mechanisms to overcome
information problems. The consumers high discount rate, however, might indeed blur
such justifications. This high discount rate suggests that there is insufficient demand in
the market for durability, since the future is worth less to the consumer. Therefore,
75
Adverse selection problem characterizes one variant of this reluctance to pay for durability. This may
happen when the durability of the high-durability goods is unobserved and varies among different looklike identical goods.
-42-
abbreviated.
(a) The Light-Bulb Industry. The incandescent-lamp cartel, which is used in this
paper to illustrate the durapolist puzzle, is probably the most studied case of technology
suppression.76
Incandescent bulbs have a simple technology which presents a tradeoff between
the bulbs lifetime and its efficiency. This tradeoff is determined by the choice of the
76
See: Stocking & Watkins [1946], at 304-362; Bright [1949] (the most thorough study of the industry and
the cartel strategies); Prais [1974]; Avinger [1981]; Reich [1992].
-43-
filaments materials and its specific structure. Essentially, the more efficiency (in terms of
light and electricity costs) the shorter the lifetime of the bulb. Therefore, the superior
product in this case is found by solving an optimization problem with few parameters
which has been always known to the bulbs producers.77
There is compelling evidence that this optimization problem has been
hypothetically solved by the industry leaders, but in their production lines they had
persisted in optimizing their profits, rather than the bulbs lifetime.78 As a consequence,
for long decades until the 1950s, the lifetime of the produced bulbs was shorter than the
optimal in fifty percent and more.79
cartelized and essentially was functioning as under the rule of one firm.80 The world
industry was organized under an international cartel, Phoebus, and domestic markets were
77
For a formal presentation of this optimization problem and its solution see: Prais [1974].
See, for example, Edwards [1944], at 16:
Protected from foreign competition, American manufacturers have attempted to increase their
business by reducing the durability of their light bulbs. They were able to find an excuse for
their action in the fact that the efficiency of a light bulb is in inverse ratio to its length of life;
but they were careful not to inform consumers of the decision to sacrifice durability, nor to
allow an opportunity for the buyer to choose between more efficient and more durable bulbs.
79
In one of the antitrust cases against General Electric, which dominated the industry, the District Court of
New Jersey quoted the following inter-departmental memorandum which was written to executives of
General Electric:
Two or three years ago we proposed a reduction in the life of flashlight lamps from the old
basis on which one lamp was supposed to outlast three batteries, to a point where the life of
the lamp and the life of the battery under service conditions would be approximately equal.
Sometime ago, the battery manufacturers went part way with us on this and accepted lamps of
two battery lives instead of three. This has worked out very satisfactorily
We have been continuing our studies and efforts to bring about the use of one battery life
lamps If this were done, we estimate that it would result in increasing our flashlight
business approximately 60 per cent. We can see no logical reason either from our standpoint
or that of the battery manufacturer why such a change should not be made at this time. United
States v. General Electric, 82 F. Supp. 753, 896-897 (D. N.J. 1949).
80
The enforcement of the cartel discipline has been a object for study of cartels stabilization. See: Priest
[1977], at 340-349; Kaplow [1984], at 1855-1862; Ayres [1987], at 318. For durable-goods cartels, see
subsection C.5.2.
78
-44-
generally dominated by single firms through their ownership of basic patents and their
licensing arrangements with other producers. Within this international scheme, General
Electric had a major role,81 and the success of the government in the 1940s in a series of
antitrust suits had virtually ended the happy days of the cartel and led to a change in the
technological evolution of the industry.82
(b) The Razor-Blade Industry. King C. Gillette invented the disposal blades, by
which his company, Gillette, gained and has maintained a global leadership (as its
executives like to state). The incipience of Gillettes blades was in the understanding that
reducing durability is often associated with profits.
King Gillette invented the disposal blades following his boss advice: King,
youre always thinking and inventing: why dont you try to invent something which
when used once, is thrown away and the customer comes back for more?83 Disposal
blades replaced the traditional, durable straight razor, and there is no doubt that they made
a significant contribution to mens daily lives.
[Shaving with razor blades,] could be performed with almost reckless boldness, as one
cannot cut himself, and in fact had become a pleasant amusement instead of an irksome
task.84
81
General Electric has never been an official member of Phoebus, nevertheless, it has operated in essential
harmony with it through foreign subsidiaries and through a long series of licensing agreements with the
leading foreign companies, financial investments in them, or both. Arthur [1949], at 307-308.
82
See: Arthur [1949], at 290-302; United States v. General Electric, 82 F. Supp. 753 (D. N.J. 1949).
83
Adams [1978], at 19.
84
Ibid, at 13.
-45-
85
86
Ibid, at 240.
Another product that Gillette favored less, despite its relatively low durability, is the disposable razor.
The disadvantage of disposable razors is that in every purchase the consumer makes a costless decision of
which razor to buy. In contrast, when the consumer has sunk costs in the razor, a decision not to buy the
compatible blades is costly. For this reason, disposable razors are regarded as Commodity Hell and
inferior to durable razors with disposable blades. See section E (tying arrangements). Gillettes decision
not to enter the market of disposable razors caused it yet another painful loss when Bic almost swept the
market in the 1970s. Gillette responded to this entry (in its European markets) with a counter-entry, as it
did in the case of Wilkinson. See: McKibben [1998], at 96-99; James Surowiecki, The Billion-Dollar
Blade, NEW YORKER, Jun. 15, 1998, at 43:
Commodity hell is exactly what American Corporations fear most, because when your
product is a commodity essentially interchangeable with all its competitors the only way to
get market share is to cut the price. That means that your profit margin is continually
dropping.
-46-
profits,87 and in the history of Gillette, this decision is perceived as a mistaken business
judgement which led to painful losses. Antitrust laws are not designated to replace
business judgement, even if this judgement brings about losses for the deciding firms.
Yet, Gillettes mistaken decision is ascribed to the period of time beginning in 1956, when
the appearance of the stainless blades could have been anticipated, and not to the entire
period since it made the decision to abandon the stainless blades.
Therefore, it is suggested, some antitrust attention should have been drawn to
Gillettes conduct in this period of time and especially to its relationships with the fringe
firms.
The case of the stainless blades illustrates that a durapolist who sticks to a low-
durability technology might eventually lose its position to fringe firms with superior
products. This, however, may occur only if the durapolist does not stifle the developments
of the fringe firms, as might be the case of the stainless blades. Gillettes patents on the
stainless-blade technology were quite broad and even Wilkinson Sword was forced to pay
a royalty to Gillette because its blade technology was covered by Gillettes patents. It is
plausible, therefore, that Gillette, which virtually did not market stainless blades, fought
against commercial attempts to introduce stainless blades through its intellectual property
rights on this technology and by other means.88
87
-47-
89
Johnson & Johnson v. McDonald, 537 F. Supp. 1282 (D. Minn. 1982), revd, 722 F.2d 1370 (8th Cir.
1983). See also: Alling v. Universal Manufacturing Corp., 7 Cal. Rptr. 2d 718 (Ct. App. 1992)
(fluorescent technology).
90
Francesca Lunzer, No Pain, No Gain, FORBES, Nov. 21, 1983, at 321. See also: Anonymous, New Killers
of Chronic Pain, BUS. WEEK, Dec. 10, 1979, at 127.
-48-
[of technology does not meet] traditional tests of establishing a per se violation of restraint
of trade under section 1 [of the Sherman Act].91
This case of the pain-control device involved financial and marketing
complications, but again its alleged scheme seems similar to the previous cases
suppression of a durable technology in the interest of higher profits.
(d) Monsantos Terminator Seeds.
largest seed corporation, acquired Delta & Pipe Land, a large American cotton-seed
company. Two months earlier, Delta & Pipe Land and the USDA were awarded a patent
on a technique that genetically disables a seeds ability to germinate when planted a
second season.92
Some observers claim that the new seeds are designed to abolish the ancient
practice of farmers to save seeds from one growing season and plant them in another
season. Simply put, it is argued that the durability of the new seeds is lower in order to
force farmers to buy seeds more often.93
91
McDonald v. Johnson & Johnson, 722 F.2d 1370, 1373 (8th Cir. 1983).
The patent covers various types of corps and not only cotton and it was awarded to both Delta & Pipe
Land and the US Department of Agriculture since the USDA invested in the research. At the time of this
writing, Monsantos acquisition is still under regulatory scrutiny. See: Tokar [1998]; Ferrara [1998];
Karen Lundegaard, Ga. Growers Sue Maker of Cotton Seeds, WALL ST. J. (interactive ed.), Mar. 24,
1999;
Leora
Broydo,
A
Seedy
Business,
MOTHER
JONES,
Apr.
7,
1998
(<http://www.motherjones.com/news_wire/broydo.html>).
93
According to observers the goal of reducing the durability of Monsantos seeds is not new and Monsanto
has tried to achieve this purpose by contractual means. Thus, it is reported, that Monsanto requires
farmers
to
sign
agreements
promising
not
to
plant
seeds
that
their
crops
produce. These agreements, of course, are inferior to terminator seeds since it is too costly to enforce
them. See: Jeffrey Kluger, The Suicide Seeds, TIME, Feb. 1, 1999, at 44; BARNABY J. FEDER, Plant
Sterility Research Inflames Debate on Biotechnologys Role in Farming, N.Y. TIMES (interactive ed.),
Apr. 19, 1999.
92
-49-
At the time of this writing, the new seeds have not been commercially marketed
yet and, therefore, their impact could not be estimated. Nevertheless, the acquisition of
Delta & Pipe Land by Monsanto represents more than a merger between two firms in a
concentrated industry, and attention should be drawn to the presumable durability
aspects.94
D.4.
94
95
If the consumer is willing to pay a premium for the new model, this
The concerns of the Department of Justice seem to focus on the vertical aspects of the merger and, in
particular, the interrelations between the development of crops and their production. Durability may be
one facet of these concerns. See: Jerry Guidera, Justice Department Picks on More than Cotton in
Monsanto Bid, DOW JONES NEWSWIRES, Jul. 1, 1999; David Barboza, Monsanto Faces Growing
Skepticism on Two Fronts, N.Y. TIMES (interactive ed.), Aug. 5, 1999.
Waldman [1993], for example, described reduced built-in durability as the cure for the durapolist
problem and reduced variable durability as the illness (the durapolist problem.)
-50-
introduction essentially decreases the value of the previous model(s). Therefore, if the
consumer anticipates such an introduction, the initial price she is willing to pay for the
new car is bounded.
presumably because a car durapolist may find it hard to resist the temptation of
introducing a new model. Past sales shrink the residual demand for new cars with which
the durapolist is faced and, hence, the durapolist has incentives to expand the residual
demand by offering new cars with slight, fashionable changes. In short, reduced variable
durability presumably represents the durapolist who does not internalize the effects of his
present behavior on the value of the durables he previously sold, and as a result may
render some inefficiencies or profit reduction.96
Inconsistent behavior of the durapolist, however, may be only a partial explanation
for the practice of frequent introductions. Considering the prosperity of durapolists who
employ this practice, their inconsistent behavior appears to be negligible in this context.
Indeed, the standard explanations for this practice attribute it to the very consistent
behavior of profit maximization.
The consistent behavior hypothesis furnishes two types of explanations for profitmaximization behavior which do not coincide with each other. These two are strategic
behavior and economies of scale, and both rely on the premise that higher frequency of
introductions of new models requires from actual and potential competitors more capital to
compete effectively.
requirements than big firms since their costs per unit are higher. Small firms costs per
96
For representative analyses of this argument see: Miller [1974]; Waldman [1993].
-51-
unit are higher because they can spread the costs of the design, engineering, retooling and
advertising only over a small volume of production.97
As a consequence, frequent
introductions are more costly for small firms than for big firms.
The explanations of strategic behavior, such as predatory conduct and raising
rivals costs, maintain that the practice is used to exclude competition and sometimes even
involves losses in the short run.98 In contrast, many commentators assert that the practice
ought to be viewed merely as a legitimate and desirable use of scale economies.99
It seems, therefore, that although time-inconsistency may urge durapolists to
accelerate the frequency of introductions of new models, it is not a central motive for this
practice.
The prevailing use of this practice by highly profitable firms, such as the
automobile manufacturers in their glamour days, presumably indicates that the influence
97
98
99
This argument is virtually a mirror image of the argument which links marginal costs to durability. If the
marginal costs are higher regardless of the durability, let say because of a tax per unit, the durapolists
monopolistic price for a one-year bulb is higher than one-third of the monopolistic price of a three-year
bulb since his markup will be significantly lower. As a result he will be induced to offer three-year bulbs
(or longer cigarettes instead of shorter ones). See: Barzel [1976] (analysis of tax per unit); Pashigian,
Bowen & Gould [1995] (analysis of the impact of increase in the costs of introducing new cars on the
frequency of introductions).
The latter logic led several economists to suggest that a durapolist who keeps his marginal costs high is
also unlikely reduce prices and, therefore, his time-inconsistency is mitigated. Namely, high marginal
costs as a device to overcome the durapolist problem. See, for example: Bulow [1982]; Kahn [1986];
Karp & Perloff [1996].
See: Menge [1962]; Yales note [1971]; Bartkus [1976]; Ordover & Willig [1981]; Ordover, Sykes &
Willig [1983]; Sidak [1983]; Petty [1988]; Kwoka [1993].
See, for instance: McGee [1973] (the study was sponsored by Ford); Selander [1973] (the study was
conducted during a clerkship for General Motors); Pashigian, Bowen & Gould [1995]; Scherer [1996], at
291-295; Areeda & Turner [1975], at 732:
The design, engineering, and retooling costs of any particular variation will obviously be less
costly per unit when spread over a larger volume of production
Yet the disadvantages from which small rivals may suffer do not warrant antitrust attack on
the monopolist who exploits them. The small-volume producer has the option of maintaining
an unchanged product while offering buyers the benefit of the costs saved by avoiding
changeovers. If the cost savings are negligible as compared to costs of the large producer who
does changeover, the small rival is simply the victim of economies of scale.
-52-
of the consistent behavior hypothesis is the decisive one. Yet, one joint underlying
assumption of the time-inconsistency and the consistent behavior hypotheses must be
examined. The two hypotheses, on their different variants, premise that the consumer is
willing to pay a premium for a new style. In many instances new styles include only slight
changes, which sometimes are merely cosmetic and insignificant and, hence, the
willingness to pay renders losses for the consumer since it presents a decline of the
opportunity costs of holding a used durable (or buying such) when a new style is
introduced.100
consumer willing to pay a premium for a new style and thereby render a loss of value to
her old goods or the good she has just bought? Or, to put it differently, why is the
consumer willing to pay more for a lower durability?101
One explanation is that the consumers simply like new goods and even small
changes are worth a lot for them. Other explanations may suggest that the consumer
choice is impaired because of various market failures.
discussed below: (a) imperfect information and bounded rationality; (b) a free-riding
problem; (c) network externalities; and (d) bounded freedom of choice in consumption.
(a) Imperfect Information and Bounded Rationality. The simplest (and the most
controversial) explanation, for the consumers willingness to pay a premium for
100
Pashigian, Bowen & Gould [1995] provided statistical evidence to the correlation between loss of value
of used durables (cars) and introductions of new durables. One aspect of this evidence is that when the
changes between two models seem to be less significant the depreciation of used goods is lower. See
also: Lazear [1986] (dynamics of downward price course of new goods).
101
In section B, it was assumed that generally the consumer was willing to pay more for a higher durability.
It was also discussed that when the discount rate of the consumer was high, she was willing to pay less for
-53-
insignificant product variations (or lower durability) is that she does not rationalize her
decisions. Quantifying the value of deluxe seatbelts, power steering, or alternatively,
renovations of the ninth edition of the Durapolists Handbook is not an easy task and it is
affected by many factors, such as advertising, fashion trends and misconceptions. This
reason is one of the main explanations for the success of the practice of the three-year
cycle in the American automobile industry which had lasted for almost fifty years until the
1970s and is discussed below.
(b) A Free-Riding Problem. Unlike the durapolist, who is a single producer,
consumers are many and dispersed. Given their collective-action problem, each consumer
takes the prices of all goods (used, new, and prospective used) as given due to her
insignificant impact on them. Consequently, each consumer sees in front of her the
positive value in purchasing a new model coupled with the negligible loss of value of her
old durable. Alternatively, the consumer does not grasp that her preference of a newmodel durable over a used, old-model durable virtually renders a loss of value of the
durable she has just purchased.
(c) Network Externalities. Network effects are the value of a good for its owner
that is ascribed to the number of users of the same good.102 Correspondingly, network
externalities are the marginal value for present consumers that is brought about by a new
consumer. This value, as discussed below, can be either positive or negative. When it is
positive, the consumer is supposedly unlikely to switch to another good and
additional durability. Yet, even when the durability was infinite the consumer was not willing to pay less
for higher durability, but rather the same price for any durability.
-54-
correspondingly, when it is negative the consumer is likely to switch. Both positive and
negative network externalities may provide another explanation for the consumers
willingness to pay for insignificant style changes.
(c1) Positive Network Externalities. Additional consumers of specific model car
spur the demand for complementary goods, such as garage services and replacement parts,
and as a result reduce the prices of these goods. Durable-goods producers sometimes use
these effects to pull consumers to new products by an asymmetric compatibility. In this
context, asymmetric compatibility means a backward compatibility of new models with no
forward compatibility of older models.
In the presence of asymmetric compatibility, users of a new model enjoy the
network externalities caused by the users of both the old model(s) and the new one, while
users of the old model(s) do lose some of the network effects when one of them switches
to a new model. Thus, the introduction of a new product, which is characterized with an
asymmetric compatibility, may render a depreciation in the old goods and willingness to
pay for the new product.
Classic examples for asymmetric compatibility are the file formats of Microsofts
Word and Excel. Old versions of Word and Excel cannot read the files of newer versions,
while the new versions can read and convert the files of older versions. Furthermore, a
user of a new version can convert her files to the formats of old versions and thereby she is
not discouraged from buying a new version.
102
See generally: Liebowitz & Margolis [1998]; Shapiro & Varian [1999].
-55-
negative externalities in different contexts and especially in those which are related to
fashion. Physical similarity, as opposed to functional similarity, is not favored by fashion
fans. From their point of view, the similarity is a negative network externality which dims
their quality signals (being in and unique.)
style may temporarily offset these types of negative externalities, while keeping the
positive ones, which are attributed to functional similarities. Thus, for example, paying a
premium for a car with an exceptional color, decorations and leather may offset the
negative externalities of similarity without losing the positive externalities which are
presented in garage services and replacement parts. This premium reflects the consumers
willingness to pay for insignificant changes, in order to signal her quality.103
(d) Bounded Freedom of Choice. Students do not choose their textbooks although
they pay for them. In contrast, their professors who choose the textbooks for them get the
textbooks for free and their costs are merely the revision of their class notes.
The
professors, therefore, do not internalize the full costs of their choice of textbooks. The
alternative may be favorable. Persistent instruction with old editions decreases the costs to
students because of the robust market for secondhand textbooks. In general, whenever the
one who dictates the standard of the consumed durables does not internalize the full costs
of her decision, a costly shift to a slightly changed model of durables may occur.
103
For early economic writing on fashion demand, see Leibentsein [1950]; Robinson [1961]. For recent
studies see: Pesendorfer [1995]; Bagwell & Bernheim [1996]. The sociologist Georg Simmel articulated
(implicitly) this combination of positive and negative externalities regarding fashion:
[F]ashion expresses and at the same time emphasizes the tendency toward equalization and
individualization, and the desire for imitation and conspicuousness. (Simmel [1904], at 308).
-56-
imperfections the revealed consumers preferences are distorted and, therefore, a desirable
104
-57-
antitrust policy should take these imperfections into account. Indeed, antitrust agencies or
courts should not determine which technology or model is superior or better for the
consumer; however, as in other contexts, they should acknowledge that the revealed
preferences may be distorted and, therefore, to direct their remedies toward these market
imperfections. Essentially, this approach is the standard antitrust view, as opposed to any
other suggestion.106
-58-
(a) The Automobile Industry. As mentioned, most of the writing in the field of
reduced variable durability has concentrated on the American automobile industry
between the 1920s and the 1970s. In this period of time, the industry was controlled by
the Big Three (General Motors, Ford and Chrysler), who employed various
anticompetitive practices such as tying and vertical arrangements.107 In addition, the Big
Three practiced a reduced variable strategy of a three-year cycle. A cycle began with the
introduction of a new design, followed by a minor facelift in the second year, and ended
with a major facelift in the third year.108 These cycles provided commentators with an
abundant field of debate over the nature of the practice.
Indeed, many commentators have produced appealing arguments, which were
usually very detailed as the complexity of an industry of multi-production-lines provides.
Nevertheless, reality has provided the simplest, but yet the most compelling evidence.
Under competition, the frequency of introductions of new models is lower, while the
quality of the products increases faster.109
107
During this period of time General Motors held on average a market share of 46%. In 1969, for example,
ninety-seven percent of domestic production of passenger vehicles was held by the Big Three (General
Motors 53.7%; Ford 26.3%; Chrysler 16.9%). The Big Three had also accounted for about ninety
percent of automobile sales in the United States between 1935 to 1970. STANDARD & POOR, INDUSTRY
SURVEYS: Autos Basic Analysis (Oct. 1, 1970), at A161. For discussions, see: Menge [1962]; Fisher,
Griliches & Kaysen [1962]; Yales Note [1971]; Selander [1973]; Kwoka [1993]; Pashigian, Bowen &
Gould [1995]; Baker [1995]; Scherer [1996], at 279-335.
108
Scherer [1996], at 291-295.
109
Scherer & Ross [1990], at 254-256; Scherer [1996], id; Baker [1995]. For the face of the industry today
see, for example: Fara Warner and Gregory L. White, Ford Bets on Safety, not Style, for Comeback of the
Taurus, WALL ST. J. (interactive ed.), Mar. 30, 1999:
People tell me the Camry is plain vanilla, says Don Esmond, general manager of Toyotas
U.S. sales unit. The Camry has been the top-selling sedan in the U.S. for two years running.
I tell them that vanilla is still the No. 1-selling ice cream. Mr. Esmond says that
-59-
introductions of new model cars fell significantly, while the quality of cars in terms of
safety, energy consumption and air pollution had generally increased.110 Since the 1980s,
the automobile manufacturers introduce new models usually twice a decade, presumably
because of the cutthroat competition.111
In short, while the cause of accelerated introductions of new cars is perhaps vague,
it might be a mystery why they do not take place under competitive conditions as the
theoretical analysis predicts.112
everywhere, you probably knew that the fashionable colors were green and yellow. After
all, intelligent readers are always acquainted with the fashions order. However, did you
know in January 1999 that the color of the beginning of the new millennium would be
blue? If not, you would probably be interested to know that the Color Marketing Group
(CMG) forecasted it accurately.113
consumers in the mid-size car market rank safety features ninth among the top 10 reasons for
choosing their car. Leading the list? Reliability and dependability.
110
Obviously, the energy crisis and of the 1970s and other macroeconomic factors have significanlty
contributed a lot to this trend. There is also some evidence that the costs of introducing new models have
increased and therefore the practice was less profitable. Pashigian, Bowen & Gould [1995].
111
See: Keith Bradsher, Car Makers Get Strong Start To 1999 Sales, N.Y. TIMES, Feb. 3, 1999, at C1;
Anonymous, A New Kind of Car Company, ECONOMIST, May 9, 1998, at 61; Anonymous, Could it
Happen Again?, ECONOMIST, Feb. 20, 1999, at 19.
112
One explanation is that competitive conditions in industries which are characterized with fashionable
goods, render monopolistic competition with differentiated goods. This sort of competition puts pressure
on producers to maintain high prices and therefore, their incentives to introduce new goods are lower.
See: Pesendorfer [1995].
113
Daniel M. Gold, Blue M&Ms were Just the Beginning, N.Y. TIMES (interactive ed.), Dec. 27, 1998; Ben
Pappas, Out with the Yellow, in with the Blue, 163 FORBES 47, Jan. 11, 1999.
-60-
114
Garry Trudeau, Hues You Can Use: Will That Taupe Fabric Still be Hot Next Year? Stand by. The Color
Cartel is on the Job, TIME, Jan. 27, 1997, at 67:
[The] Color Marketing Group [is] the Virginia-based color cartel that
has held a largely unknowing public under its sway for more than 30 years. It
was the CMG that forecast avocado refrigerators in the late 60s and mauve motel
rooms in the 70s and hunter-green automobiles in the 90s. And it was the CMG
that predicted the 1996 consumer palette would be, in the words of former
president Laraine Turner, kissed by the yellow.
115
Ibid.
116
According to CMGs publications, it is a not-for profit, international association of 1,500 Color
Designers involved in the use of color as it applies to the profitable marketing of goods and services.
117
CMG hosts several leading competitors from the automobile industry, such as Ford, General Motors,
Mercedes-Benz, Honda, Toyota, Volkswagen, and others. Likewise, competitors like Adidas, Nike, and
Reebok also find it useful to forecast together the future colors. Indeed, this appeal of joint forecasting
seems to attract competitors from many industries to CMG.
118
Deborah L. Jacobs, The Titans of Tint Make Their Picks, N.Y. TIMES, May 29, 1994, at C7. See also:
Daniel Akst, The Culture of Money: Having Our Colors Done, N.Y. TIMES, Apr. 4, 1999, at C6:
-61-
predictions by the media, which assures that manufacturers will follow them.119 It is not
surprising, therefore, that some observers have defined the mission of CMGs
conferences as [determining] which colors will be big sellers on the nations clothing
racks, showroom floors and supermarket shelves.120 Or, as the slogan of CMG goes:
Forecasting the Color of Profit.
To be sure, the coordination of future fashionable colors is essential for durapolists
and their weaker relatives, the durable-goods producers, in employing costlessly reduced
variable durability.
retooling of the production lines nor an increase in the advertising budget. However, such
a change does not make the durables of the last season obsolete as long as their colors are
still sold on the market for the same price. Therefore, a coordination of all the central
fashion leaders is required in order to use colors as a device to make durables obsolete.
CMG presumably facilitates such coordination.121
D.5.
The Mount Olympus of the world of color is an organization called the Color Marketing
Group, in which people from various industries get together either to divine or dictate,
depending on our level of cynicism, the colors that people will like and want.
119
It should be noted that CMG is not the only player in the color industry, although it seems to dominate
it, considering that its competitors are few and smaller. In general, however, it will not be farfetched to
state that the industry is not characterized with competition.
One of the main competitors of CMG, is the Color Association of the United States (Caus), which was
established over eighty-five years ago. This association, according to one of its executive directors
use[s] a little art, a little science, some economics, some psychology [and] some fortune-telling in order
to forecast the next fashionable colors. See: Anastasia Toufexis, The Bluing of America: Color
psychology is used routinely to manipulate tastes, TIME, Jul. 18, 1983, at 62; Lisa Belkin, The In Colors
of 87 are being Picked in 85, N.Y. TIMES, Dec. 26, 1985, at A1. For the color industry in the United
Kingdom, see: Stephen Hancock, The Colours of Money, MARKETING, Sep. 28, 1995, at 42.
120
Jacobs, supra note 118.
-62-
The distinction between built-in durability and variable durability is often blurred
and so is the question whether the durability was reduced. This blurring makes the task of
dealing with planned obsolescence even harder. Nevertheless, these difficulties only stress
the significance of antitrust treatment. As discussed, imperfect information has never been
an exception to the antitrust laws, albeit lack of information limits judiciary and regulatory
scrutiny to a large extent. Imperfect information, as above mentioned blurring represents,
forestalls the existence of healthy competition in the markets and, thus, warrants antitrust
intervention.122 This subsection discusses circumstances in which there might be some
uncertainty.
Elimination of secondhand markets is a hybrid type of planned obsolescence. This
strategy has the characteristics of both reduced built-in durability and reduced variable
durability. Simply put, elimination of secondhand markets cripples the free market and in
this sense it constitutes a restraint of trade and renders misallocation of resources.123 The
121
For an interesting theoretical analysis, which predicts fashion cycles in colors, see Karni & Schmeidler
[1990].
122
See: Beales, Craswell & Salop (1981); Craswell (1982); Eastman Kodak Co. v. Image Technical Services,
Inc., 504 U.S. 451, 475-476 (1992):
Given the potentially high costs of information and the possibility that a seller may be able to
price discriminate between knowledgeable and unsophisticated consumers, it makes little
sense to assume, in the absence of any evidentiary support, that equipment-purchasing
decisions are based on an accurate assessment.
123
See: Salop Scheffman [1987]. The inefficiencies of the dead hand rule may illustrate the costs of
crippling the secondhand markets. The dead hand rule refers to a situation that the former owner of a
good (seller, legator, etc.) restricts the use of the good by the present owner. See: Posner [1998], 18.3.
The Widener library at Harvard expounds upon these inefficiencies (although not a restraint of trade).
The library was donated in the memory of Harry Elkins Widener, a known bibliophile, who went down
with the Titanic. When she made the donation, Wideners mother stipulated that the library could not be
remodeled (not a brick, stone, or piece of mortar shall be changed). As a result, in order to build a
breezeway between the Widener and Houghton Libraries, the architects had to run it out the window to do
it legally.
-63-
existence of the strategy emphasizes the failure of the theory of the durapolists indirect
control over the secondary market.124
Elimination of secondhand markets reduces durability because it delimits the
lifetime of a good to the use of its first user, although subsequent users could have used it
as well. This reduced durability can be determined both before the durables were released
into the market and afterwards. For example, the durapolist can condition the sale with
resale restrictions,125 or after the durables are on the markets the durapolist can recapture
them for elimination.126
As already mentioned, elimination of secondhand markets has anticompetitive
effects other than overcoming the durapolist problem. Usually, this practice is used to
handicap rebuilders of equipment and, thereby, to prevent an alternative, cheap source of
competing supply of competing durables or maintenance services.
124
Elimination of
This theory was originated by Judge Lernad in the Alcoa case. See subsection C.1.
See, for example: Williamsburg Wax Museum, Inc., v. Historic Figures, Inc., 810 F.2d 243 (D.C. Cir.
1987) (the alleged wax-figures durapolist imposed resale restrictions).
126
See, for example: Pacific Mailing Equipment Corp. v. Pitney Bowes, Inc., 499 F. Supp. 108 (N.D. Cal.
1980) (destroying mailing machines in order to eliminate secondhand markets; the court prohibited the
practice); Allen-Myland v. International Business Machines Corp., 33 F.3d 194 (3rd Cir. 1994), cert.
denied, 513 U.S. 1066 (1994) (IBM recaptured old parts in order to eliminate the secondhand markets);
United States v. Varian Associates, 1992-1 Trade Cas. (CCH) 69,772 (N.D. Ill. 1992) (consent decree
prohibiting defendant from acquiring used equipment).
The practice of eliminating the secondhand market is especially popular in the textbook industry.
Textbook publishers shred substantial numbers of books to keep them from competing with their titles at
full retail prices. They hire companies (destruction companies) to remove and shred unwanted books.
McGraw-Hill, the largest textbook publisher in the US, filed in 1994 a suit against 20 used-book dealers
and destruction companies for reselling used McGraw-Hill textbooks scheduled for destruction.
McGraw-Hill stated that textbooks valued at $20 million were diverted from recycling shredders from
October 1993 to December 1994 and that 60 percent reached the market. Jon Nordheimer, Pulp
Nonfiction: When Textbooks Refuse to Die, N.Y. TIMES, Aug. 14, 1995, at D7; Anonymous, McGraw-Hill
Settles Suit on Book Reselling, N.Y. TIMES, Jun. 18, 1996; at D9. It should be noted that in other
instances books are destroyed in order to lower the inventory costs.
125
-64-
secondhand goods was condemned by several courts for this reason, 127 while one court
also acknowledged the connection between time consistency and such elimination.128
Given these reasons for elimination of secondhand markets, the only plausible
justification for the practice may be the initial investment in creating the markets (or
developing them). It is often argued that the ability of an inventor (or a creator) to recoup
his investment in R&D is curtailed because of imitators who have access to his goods. In
addition, the durapolist problem implies that even without imitators a gifted inventor may
be curtailed as well, because of his time-inconsistency problem. Therefore, under these
circumstances, the practice should allowed some immunity from the antitrust laws. In
general, however, absent this defense, the durapolists practices that are designed to
eliminate the secondhand market should be considered a violation of 2 of the Sherman
Act.
D.6.
127
See, for example: Varian Associates (1992); Allen-Myland (1994), at 203 n. 11:
By recapturing old parts from upgraded mainframes, IBM effectively curtailed [its
competitors] ability to reconfigure their used machines into different models that could have
competed against IBMs [models].
128
Id, at 203:
[Alcoas analysis of time consistency] is particularly true when, as here, that control was
enhanced by IBMs policy of recapturing old parts that could otherwise have been used to
extend the useful service lives of existing used mainframes by allowing them to be upgraded
-65-
therefore, if the antitrust laws are designed to address the latter practices they should also
address planned obsolescence.
The doubts regarding such a treatment, I believe, deviate from the standard
antitrust approach for no apparent serious reason. Undoubtedly, durability is one of the
parameters, as price and quality, which define a product. Therefore, if durapolists and
durables manufacturers employ anticompetitive practices to set durability in order to
increase their profits their practice should be treated by courts as price fixing or quality
fixing, which are generally illegal per se. For example, if car manufacturers fix the colors
for the next season it is not substantially different from fixing prices.
Likewise,
Consumers may be willing to pay for presumably insignificant attributes because of the
same market imperfections that enable monopolists to exploit more profits through tying.
On the same token, reducing durability in order to force the consumer to buy more often is
substantially equivalent to tying, which dictates to the consumer the goods she buys and
the terms of the transaction. In short, when a durapolist acquires his market power and
maintains it through the business acumen of reducing durability, there is no preliminary
reason not to infer that a durapolist violated 2 of the Sherman Act.
Again, this
conclusion does not contradict the existence of legitimate justifications for overcoming the
durapolist problem as discussed.
and placed with new customers. We therefore conclude that the district court erred when it
added leases of used IBM mainframes into the relevant market.
-66-
E.
E.1.
Tying Durables
A tying arrangement is a sale (or lease) of one product or service on condition that
the buyer (or the lessee) takes another product or service.129 Durability is a prevailing
characteristic of tying goods,130 and this characteristic seems also to be in the minds of
lawyers and economists when analyzing tying arrangements.
accompany the tying analysis are of a machine to which are tied complementary goods
(such as replacement parts and services) or production inputs (such as toners and paper).
Tying goods, however, are not exclusively durables, and as the definition of the tying
arrangement provides the tying good might be a consumable as well.131
Tying arrangements may be condemned under the antitrust laws or under the
intellectual property laws, where intellectual-property rights are involved.132 Roughly
speaking, the legality of tying arrangements depends to a large extent on the market power
in the tying market with some extensions usually related to the anticompetitive effects in
129
See: Northern Pacific Railway v. United States, 356 U.S. 1, 5-6 (1958); Scherer & Ross [1990], at 565569.
130
See: Areeda, Elhauge & Hovenkamp [1996], at 1732b:
The tying product has usually been more durable than the tied product. The typical tying
product was a machine that processed, handled or consumed the tied product.
131
See for example: Jefferson Parish Hospital No. 2 v. Hyde, 466 U.S. 2 (1984) (hospital services); TimesPicayune Pub. v. United States, 345 U.S. 594 (1953) (advertising space).
132
See: Areeda, Elhauge & Hovenkamp [1996], 17; Hovenkamp [1994], 5.5, 10. Under the federal
antitrust laws, tying may be illegal under 1-2 of the Sherman Act (15 U.S.C.A 1-2), 3 of the
Clayton Act (15 U.S.C.A. 14) and 5 of the FTC act (15 U.S.C.A 52). When intellectual property
rights are involved, tying arrangements may be condemned under the Patent Misuse Doctrine and its
extensions to other intellectual property rights. The 1988 Patent Misuse Reform Act (35 U.S.C.A.
271(d)) narrowed this doctrine to a large extent. See also: The 1995 Antitrust Guidelines for the
Licensing of Intellectual Property.
-67-
the tied markets. This law of tying seems to have provided a perpetual battlefield for
antitrust commentators on the question of whether tying may facilitate an extension of
market power.133 Perhaps because of this persistent war, the durapolist problem has not
been treated in the context of tying. This section addresses the tying arrangements as a
device to overcome the durapolist problem. In doing so, I join those who argue that tying
may extend market power.
E.2.
characterized by longevity. Potentially, there are 100 customers that are divided into k
133
There are two fundamental views regarding this issue: the efficiency-motive view and the market-powermotive view. In general, the efficiency-motive view, which is associated with the Chicago School, argues
that tying is a device for maximizing profits in a given market structure. See: Director & Levi [1956];
Bowman [1957]; Posner [1976], at 171-184; Bork [1978], at 140-144, 365-381. In contrast, the marketpower-motive view argues that tying may also serve as a device to extract more profits through market
distortions such as barriers to entry, market foreclosure and extension of market power (leverage). See:
Turner [1958]; Kaplow [1985]; Whinston [1990]; Slade [1998]. The market-power-motive view was
originated in early tying cases. See: A. B. Dick (1912) (Chief Justice white, dissenting); Motion Picture
(1917); Carbice Corporation of America v. American Patents Development Corp., 283 U.S. 27 (1931).
-68-
customers according to the intensity of their use of the machines: $2 per year for the highvalue customers, and $1 per year for the low-value customers. The inputs markets are
perfectly competitive and the interest rate is 0.1.
Under these assumptions, the inputs prices are their costs and, therefore, if the
durapolists sells his machines, the customers are willing to pay for the machines the net
present value of their future payments ($5 or $2).
willing to pay up to $50 per machine, and the low-value customers are willing to pay up to
$20 per machine.
Three scenarios portray the durapolists profits under these assumptions: (a) no
price discrimination and no commitment; (b) price discrimination; and (c) commitment.134
Under the scenario of no price discrimination and no commitment, the durapolist
sells 100 salt-dispensing machines for $20 each, earns $2,000 and the expected social
134
The scenario of price discrimination, as we will see, provides the strongest commitment and, therefore,
when the homogeneity of the customers is relatively high, as in the example (two groups of identical
customers), there is no need of a specific commitment under a scenario of price discrimination. When
-69-
surplus is $3,500. Under the scenario of price discrimination, the durapolist is expected to
sell 50 machines for $20 each and 50 machines for $50 each. The durapolists profits
under this scenario are $3,500 and this is the entire social surplus. Finally, under the
scenario of commitment, the durapolist is expected to sell 50 machines for $50 each, to
earn $2,500 and this amount is also the entire social surplus.
This simple example suggests that absent distributive considerations, from the
social point of view, the scenario of no price discrimination and no commitment is
equivalent to the scenario of price discrimination. In contrast, the scenario of commitment
is inferior in comparison to the former scenarios because it yields a lower social surplus.
Accordingly, if price discrimination is effectively forbidden and the durapolist can commit
to future prices, society is worse off.
Therefore, supposedly there are instances in which price discrimination is better
for society. Indeed, this observation is generally accepted when R&D investments are
involved. Under these circumstances, it is largely held that price discrimination may be
justified to some extent, in order to furnish inventors and creators with the appropriate
incentives to invest in R&D activities.135 Nevertheless, the analysis of tying in the context
the consumers heterogeneity is high a commitment may put the durapolist in a better position, but
essentially does not change the analysis.
135
The Robinson-Patman Act, 15 U.S.C.A. 13, presumably prohibits price discrimination. See Hovenkamp
[1994], at 14.6. For a discussion concerning the intellectual property rights and price discrimination,
see: Bowman [1957], at 15-31; Baxter [1966]; Bowman [1973], at 98-118; Posner [1976], at 177-179;
Kaplow [1984], at 1873-1882; Scherer & Ross [1990], at 494-502. For a formal analysis, see: Husman &
Mackie-Mason [1988].
-70-
of price discrimination often fails to recognize the scenario of commitment.136 This facet
of tying is my focus in this section.
addition he earns $1,000 on his machines. Thus, the durapolists total expected profits are
the net present value of these incomes, which equals to $3,500.138
As the example implies, price discrimination is a solution to the durapolist
problem. The durapolists time-inconsistency problem stems from the perception of his
customers that he employs intertemporal price discrimination.139
However, if the
durapolist discriminates in prices, postponing purchases is unlikely to reduce the price for
the customer and, thus, the durapolist does not cope anymore with time-inconsistency.
136
Ordover, Sykes & Willig [1985] and Whinston [1990] provided models in which tying could be
interpreted also as a commitment. See, also: Borenstein, Mackie Mason & Netz [1995] (discussing the
profitability of tying in durable-goods markets).
137
Tying that facilitates price discrimination is often called metered tying. For a general analysis, see:
Hansen & Roberts [1980]; McGee [1987]. For an illustration of price discrimination through tying of
automotive replacement parts, see Scherer [1996], at 308-312.
138
139
Stokey [1979] showed why a monopolist might be worse off when he employs intertemporal price
discrimination. Essentially, her model addresses the durapolist problem.
-71-
-72-
evidence, I believe, does not preclude the viability of my analysis. Price discrimination is
a better strategy for the durapolist because he captures more surplus than under
commitment without price discrimination. Therefore, as long as price discrimination is
possible, a strategic use of tying in order to create commitment to future prices is a
dominated strategy.
E.3.
Policy Considerations
The above discussion leads to two conclusions. First, tying may facilitate price
discrimination to overcome the durapolist problem. There is one new aspect in this
proposition, which is focused on overcoming the durapolist problem. The link between
The durapolist profits on each package of inputs are $5 and, therefore, = 50 05.1 = 2 ,500 . United Shoe
Machinery had similar pricing (free services), but its contracts were for long terms and it could
discriminate between customers. See Kaysen [1956], at 250-256.
141
Again, in a certain point of time the durapolist cannot price-discriminate because of the law enforcers.
142
To some extent, the cases of United Shoe Machinery (Kaysen [1956], American Can and Continental Can
(McKie [1955; 1959]) may fit the described scenario although they also contained price discrimination.
The cases of the durable-goods sellers that succeeded in tying the aftermarkets to their goods fit to some
extent this analysis, although these cases are usually associated with various market imperfections. See,
140
-73-
tying and price discrimination has been thoroughly studied. In this sense, the analysis
illustrates how tying and price discrimination can be employed to extend market power.
Overcoming the durapolist problem through these practices seems to be a neglected
perspective.
Aside from this perspective, the standard analyses of tying for price
discrimination and the desirability of price discrimination in various settings fit the
durapolists case as well.
The second conclusion refers to tying as a device for commitment. If tying is
permissible and price discrimination is impossible, a durapolist may overcome his
problem through tying that creates a credible commitment.
This conclusion provides a new perspective of tying arrangements.
Tying
arrangements may have anticompetitive effects even if they are not used to leverage
market power to related markets or to deter entry into the tying market. Tying may enable
the durapolist to acquire more market power than he would have had otherwise.143
This perspective presents another fragile link in the theory concerning a fixed sum
of market power.144 One link has gained much publication and study since Kodak (1992).
In Kodak, the Supreme Court analyzed Kodaks argument that higher service prices will
for instance: Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992). For an analysis
of these cases, see: Borenstein, MacKie-Mason & Netz (1995).
143
The Jerrold Electronics case illustrates the emphasis on the incremental increase in the market power of
the durapolist. In this case, tying of services was held as reasonable restraint because it was instituted in
the launching of a new business with a highly uncertain future. [But as] the industry took root and grew,
the reasons for the [tying] disappeared. Namely, tying is reasonable if it constitutes a necessary
condition for the existence of the market, even it creates the monopoly. This rule is equivalent to the
discussed considerations to allow price discrimination when intellectual property rights are involved.
United States v. Jerrold Electronics Corp., 187 F.Supp. 545, 557 (E.D. Pa. 1960).
144
See, Posner [1976], at 173: [A] fatal weakness of the leverage theory is its inability to explain why a firm
with a monopoly of one product would want to monopolize complementary products as well.
-74-
lead to a disastrous drop in equipment sales.145 The majority held that the existence of
information and switching costs break the link between equipment (durables) and services
(consumables). Similarly, when tying is used to provide a credible commitment to high
prices, higher prices of consumables lead to a beneficial reduction in durables sales.146
To the extent that the federal antitrust law of tying can be characterized,147 the
perspective of tying as a device to create a credible commitment appears to be uncovered
by the antitrust law because of its stress on the tied market.148 Under the so-called per se
rule against tying, this perspective is not always covered, because of the requirement for
an involvement of a not insubstantial amount of interstate commerce in the tied product
market.149 Likewise, under the rule of reason, the courts examination is usually focused
on the market power in the tied market,150 which is not required in order to overcome the
durapolist problem through tying. Indeed, a durapolist may have market power in the
145
Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451, 472 (1992). For a another analysis
of this false dichotomy in the copier industry, see Blackstone [1975] (the case of SCM).
146
The Kodak court explained that [the] significant information and switching costs [in markets for
complex durable goods] could create a less responsive connection between service and part prices and
equipment sales (at 473). The commitment perspective implies that even a responsive connection does
not necessarily exclude anticompetitive effects.
147
This aspiration appears to be almost farfetched due to the inconsistency of the way the tying law is
applied by the courts. See: Areeda [1991], 1700-1701, 1730.
148
See, for instance, Northern Pacific Railway v. United States, 356 U.S. 1, 6 (1958):
[Tying arrangements] deny competitors free access to the market for the tied product, not
because the party imposing the tying requirements has a better product or lower price but
because of his power or leverage in another market. At the same time buyers are forced to
forego their free choice between competing products.
See, also, Jefferson Parish Hospital No. 2 v. Hyde, 466 U.S. 2, 36 (1984) (OConnor, J., concurring):
Tying may be economically harmful primarily in the rare cases where power in the market
for the tying product is used to create additional market power in the market for the tied
product.
149
See: Fortner Enterprises v. U.S. Steel Corp., 394 U.S. 495, 498-499 (1969); United States Steel Corp. v.
Fortner Enterprises, 429 U.S. 610 (1977).
150
Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992).
-75-
inputs markets either because of his size or because he created a technological tying,151 but
the existence of this market power is not a necessary condition.
In summation, tying is typically of consumables to durables, and the implications
of this setting appears to be neglected by durapolists researchers. Tying may be employed
by durapolists to overcome their time-inconsistency problem and this is the promising
opportunity that tied consumables provides. This perspective sheds a new light on the
anticompetitive effects of tying arrangements which has generally been ignored. In legal
terms, tying is a device to acquire and maintain monopoly power by durapolists in their
markets, and not only to leverage their market power to related markets.152 Again, as
emphasized throughout this paper, this presumable violation of 2 of the Sherman Act
may be justified to encourage creative activities, but the nature of these justifications goes
beyond the scope of this paper. Once the concerns for impairing the competition have
been established, as in this section, there is reason to progress and to examine these
justifications.
F.
F.1.
(hereinafter: leases) have several important economic functions such as risk allocation and
151
Technological tying usually means that the tied inputs have value only when used with the tying good.
For instance, replacement parts for a Kodak copier have no value if they are not used with such a copier.
See, generally: Areeda, Elhauge & Hovenkamp [1996], 1757.
-76-
financial devices. Leasing, however, may be used for anticompetitive purposes as well
and, therefore, lease practices and especially lease-only policies have been one of the main
study subjects in durable-goods markets.
A lease-only policy reminds us of planned obsolescence in the sense that both offer
goods for shorter periods of time than they could have been offered. For this reason,
several commentators have suggested that the analysis of a lease-only policy and planned
obsolescence is generally the same.153 Given the complexity of planned obsolescence, the
appeal of this view is that lease policies supposedly provide an aesthetic analytical
framework since they are observable and verifiable.
Indeed, as discussed below, the look-alike lucid appearance of lease policies has
tempted many teachers to focus on this facet of lease policies in addressing the durapolist
problem and his ways to get around it.154 This once popular theory is generally not
accepted out of textbooks and in this section I discuss the whys and wherefores.
Furthermore, the use of leasing practices as a device to overcome the durapolist problem
indirectly is also discussed. Leasing as already noted may facilitate other practices that
assist in overcoming the durapolist problem. In light of the rich literature on this practice,
the discussion is abbreviated.
152
-77-
F.2.
The leasing durapolist does not sell durables. His consumers essentially buy a
stream of consumables (machine services) from him. Therefore, by leasing, the durapolist
switches the goods he sells from durables to consumables. To put it differently, when the
durapolist leases his durables he allegedly internalizes the costs of an inconsistentbehavior pattern, because the durables ownership does not change and remains in the
hands of the durapolist. For this reason and others, the practice of lease-only is generally
illegal under the antitrust laws.
A simple illustration of the lease-only policy as a device to overcome the
durapolist problem can be derived from the example which was used to clarify the value of
commitment in subsection C.1.
This time, consider a copier durapolist who incurs no production costs and sees in
front of him high-value and low-value consumers. There is no discount rate, the highvalue consumers are willing to pay up to $5 a year for copier services and the low-value
consumers are only willing to pay up to $2 a year. Assume that there are 100 consumers
and k of them are high-value consumers, where k is unobservable and distributed
uniformly over [0,100]. A consumer does not need more than one copier, but if the price
is too high she will not have a copier at her office. If the copiers lifetime is ten years, the
Professor Coase has argued that the lease-only policy of a monopolist of a durable-good
may be designed to overcome the difficulties encountered in trying to charge a monopoly price
for a durable good. Perhaps the lease-only policy should have been forbidden on that ground.
-78-
durapolist can sell it to the high-value consumers for $50 and to the low-value consumers
for $20.
If the durapolist sells his copiers, the result is as discussed in section in subsection
C.1. Under a policy of non-commitment, the durapolist sells 100 copiers for $20 each and
his profits are $2,000. In contrast, under a policy of commitment, the durapolist sells k
copiers for $50 each and his profits are 50k. Thus, from the durapolists standpoint, it is
profitable to commit if there are more than 40 high-value consumers.
Given the
distribution of the consumers, the probability for more than 40 high-value consumers is
60% and the expected profits under a commitment are 2,500.
Assume, now, that at time zero, the durapolist decides between outright sale of his
copiers and leasing them for one-year periods.155 In this case, the durapolist decides to
lease at a price of $5 a copier for annual expected profits of $250, because under the
alternative of outright sale his respective profits per year are $200. If the durapolist leases,
in the second year he already knows the number of high-value consumers and probably
their identity, since they revealed themselves in the first year when they leased the copier
for $5. Accordingly, the durapolist can decide what would be the best pricing for his
copiers.
If the high-value consumers are identified, starting in the second year, the
durapolist price-discriminates between his customers, i.e., he charges the high-value
155
As already discussed, the advantage of reduced durability relies on the assumption of discrete selling or
short-run price rigidity. Absent this assumption, a reduced durability or a lease-only policy may not assist
to overcome the time-inconsistency problem. For a formal analysis of this argument in the context of
leases, see DeGraba [1994].
-79-
customers $5 a year and the low-value customer $2 a year. This way, the durapolists
profits are maximized and the deadweight loss is minimized.156
The other scenario is that the durapolist cannot price discriminate between his
customers. This can happen for three reasons: (1) the high-value consumers anticipated in
the first year the consequences of their willingness to pay and, therefore, they did not
reveal themselves; (2) k varies annually, i.e., there are fluctuations in the demand of the
high-value consumers; and (3) there are effective legal restrictions against price
discrimination.
Under this scenario, the durapolists profits are $200, when either (a) price
discrimination is prohibited and k is fixed and below 40; or (b) the high-value consumers
are strategically smart. When k varies, the durapolists expected profits are $250. And
when k above 40 and the durapolist credibly commit to his future lease policy his profits
are more than $250.
This example provides a simple lesson. High-value customers are unlikely to enter
to lease contracts at high prices, reveal their identity and thereby expose themselves to
price discrimination.157 The durapolist, therefore, may improve his position with a leaseonly policy only if he commits to a non-discrimination policy, commits to future rents, or
156
Under price discrimination, the durapolists annual profits are expected to be $350. In expectancy there
are 50 high-value customers and 50 low-value customers and their aggregate rents are 50*5+50*2=350.
The deadweight loss is minimized because all of the customers purchase copier services. For extended
discussion, see subsection E.2.
157
Note that this case does not present a collective action problem such as the case of tying. Here it is
profitable for each high-value consumer not to reveal himself regardless of what the rest of the consumers
of his type do. Cf. subsection E.2.3.
-80-
does not keep records of his customers.158 To put it differently, the success of lease-only
depends on a commitment for consistency. Lease-only policies, therefore, do not seem to
overcome time inconsistency because their alleged success assumes a commitment to
consistency.159
F.3.
The leasing durapolist was popularized among lawyers and economists by Judge
Charles Wyzanski and his law clerk, the economist Carl Kaysen, who worked together
on the United Shoe Machinery (USM) case.160 Judge Wyzanski condemned the lease-only
practice, mainly because he perceived it as a high barrier to entry to the shoe machinery
market. As a consequence, USM was obliged to sell its machines at prices which were
equivalent to the respective lease prices.
Indeed, the use of this practice in order to overcome a time-inconsistency problem
was not addressed by the court, perhaps because Kaysen was not familiar with such an
analysis that was relatively new at the time.161 Instead, the court stressed the fact that
158
The Robinson-Patman Act, 15 U.S.C.A. 13, presumably prohibits price discrimination. Hovenkamp
[1994], at 14.6. Hart & Tirole [1988] provided a formal model which showed that the durapolists
profits are lower in the rental case without any form of commitment than in the case of outright sale.
Fudenberg & Tirole [1998] provided a model which linked the information concerning the consumers
identity to solutions for time-inconsistency problems.
159
An alternative conclusion is that lease-only policy may be beneficial for the durapolist, if a prohibition
against price discrimination is enforced effectively. See subsection E.2.3.
160
United States v. United Shoe Machinery Corp., 110 F. Supp. 295 (D.Mass.1953), affd per curiam, 347
U.S. 521 (1954). Kaysens Ph.D. dissertation was based on that case and in 1956 he also published his
analysis in a book. Kaysen [1956].
161
See: Strotz [1955]. Judge Wyzanski endorsed Kaysens view that a purchased machine generally
presented a sunk cost and therefore a rational consumer would not take it into account in his calculation.
In doing so, the court did not consider that the value the consumer can save from her investment in a
-81-
USM had maintained its market power through deterring entry by long-term leases and
elimination of the secondhand market for its machines.162
The USM decision was not questioned by the Supreme Court and it has provided
the authority for successful government attacks on leasing policies in other industries and
against other giants such as IBM, Xerox, American Can and Continental Can.163 The
reasons for these successes may be worth a study due to their contrast to cases where
midgets were involved and the courts rejected the what appears to be a novel theory
under the antitrust laws.164
machine could have been decreased due to a time-inconsistency problem. See: United Shoe (1953), at
324; Kaysen [1956], at 68-69.
162
As discussed, secondhand markets constitute part of the durapolist problem, because they provide the
consumers with a competitive, cheap source of supply. Therefore, any wise durapolist would zealously
endeavor to eliminate them. Kaysen [1956], at 73; Waldman [1997] analyzed the role of a lease-only
policy as a device to eliminate the secondhand market. See: United Shoe (1953), at 325:
Uniteds lease system makes impossible a second-hand market in its own machines. This has
two effects. It prevents United from suffering that kind of competition which a second-hand
market offers. Also it prevents competitors from acquiring United machines with a view to
copying such parts of the machines as are not patented, and with a view to experimenting with
improvements without disclosing them to United.
163
The famous cases are those in the industries of computers, copiers and can machinery. These cases were
subjects for several case studies, sometimes for the purpose of litigation. For the computer industry, see:
United States v. International Business Machines Corp., 1956 Trade Cas. (CCH) 68,245 (S.D.N.Y.
1956) (consent decree); Brock [1975], at 155-159; Fisher, McGowan & Greenwood [1983], at 191-196;
DeLamarter [1986], at 106-117. For the copier industry, see: In re Xerox Corp., 86 FTC 364 (1975);
Blackstone [1975]; Breshnahan [1985]. For the can-machinery industry, see: McKie [1955, 1959].
164
Souza v. Estate of Bishop, 821 F.2d 1332, 1334 (9th Cir. 1986). In Souza, the defendant apparently
dominated the market for residential land in Honolulu. The Souza court cited in its decision the decision
of the second circuit in the case of Berkey Photo v. Eastman Kodak: [A]ny firm, even a monopolist, may
generally bring its products to market whenever and however it chooses. Berkey Photo, Inc. v. Eastman
Kodak Co., 603 F.2d 263, 286 (2d Cir.1979). In the Berkey case, Kodak allegedly employed a practice of
planned obsolescence (variable reduced durability) in order to monopolize the market. See supra note
105 and accompanying text.
In another case, a private antitrust suit was filed against an alleged wax figure durapolist. Williamsburg
Wax Museum, Inc., v. Historic Figures, Inc., 810 F.2d 243 (D.C. Cir. 1987). This durapolist refused to
sell his wax figures to one of his customers, but agreed to lease him the figures. In addition, the
durapolist imposed different resale restrictions on his customers. The plaintiff argued that these policies
were designed to preclude a secondhand market. (at 246). While the court acknowledged that [l]easing
requirements, in combination with other practices, can sometimes impede competition and thereby violate
-82-
Professor Coase might be inspired by some of these cases, although it seems that
he did not reckon that the USMs leases were used to overcome the durapolist timeinconsistency problem.165
In textbooks, however, the USM case is often used to illustrate the Coase
conjecture and the solution for time-inconsistency problems.166 This popularity has not
gained a general acceptance among lawyers and economists and many cast doubts on the
applicability of the Coase conjecture in the USM case.167 In particular, it is argued that
USM dictated long-time leases and fined its customers for terminating leasing contracts.168
Furthermore, USM had kept very detailed records of its customers, which were used to
price-discriminate against its customers. Therefore, It would be fair to say that USMs
lease-only policy has not been used to overcome any time-inconsistency problem,
the antitrust laws, it dismissed the allegation that the specific combination in that case (leasing and resale
restrictions) constituted a violation of the antitrust laws.
165
Professor Coases analysis was theoretical and he did not mention USM or any other durapolist.
Nevertheless, while the USMs leases were for ten years, Professor Coase argued that a lease-only policy
assists in overcoming time inconsistency if the durapolist leases his durables for relatively short period
of times. Coase [1972], at 145.
166
See, for example: Posner [1976] at 184; Tirole [1988], at 81-83; Scherer & Ross [1990], at 457.
167
See, for example: Bork [1978], at 136-160, 164-175; Flath [1980]; Wiley, Rasmusen & Ramseyer [1990];
Goldberg [1990]; Masten & Snyder [1993].
168
According to USMs formal policy, which was stated in its lease contracts, in the case of the replacement
of a USMs machine by a competitive machine, the lessee continued to be liable for the unexpired period.
In practice, USM did not enforce this policy, although it penalized many customers who turned to its
competitors. The choice of the penalized customers and the magnitude of their fines were unclear and
appear to be random. This combination of a formal policy with the informal looked-alike random policy
has increased, of course, the perceived costs of breaching leasing contracts with USM. Kaysen [1956], at
64-73. The data that Kaysen used for his analysis referred to the period of time after the governments
complaint against USM was filed (Dec. 15, 1947). Therefore, it is generally possible that generally the
fines and the respective perceived costs of breaching were higher.
-83-
although it had other anticompetitive effects, such as deterring entry, tying, price
discrimination and elimination of secondhand markets.169
F.4.
Policy Considerations
The widespread opinion concerning the absence of connection between the USMs
lease-only policy and overcoming its time-inconsistency problem is convincing. A leaseonly policy, however, may serve various purposes other than overcoming the time
inconsistency. In particular, it may facilitate price discrimination,170 tying,171 elimination
of secondhand markets,172 and erecting barriers to entry.173 These practices may be used
169
For a formal analysis of long-term contracts as a barrier to entry, see Aghion & Bolton [1987]. Longterm leases are equivalent in this sense to selling, but they are coupled with more advantages, such as
tying and price discrimination. Elimination of secondhand markets (subsection D.5) and tying (section E)
may be used as devices to overcome the durapolist problem. It should be noted that in the leading cases
of lease-only, one of the main goals of the practice was to prevent fringe firms from access to the
machines and thereby developing competing machines. See supra note 163.
170
IBM Corporation v. United States, 298 U.S. 131 (1936) (business machines); Kaysen [1956], at 75-78,
124-130 (Shoe Machinery); McKie [1959], at 55-65 (can-closing machinery). The Williamsburg case,
supra note 164 may also illustrate a technique of price discrimination. In Williamsburg, the alleged waxfigure durapolist, Lynch Display Corporation, fixed the rent for its wax figures on 10% of the gross
receipts of the wax museums and, thereby, the rent varied according to the income of the museums from
their exhibitions of Lynchs wax figures.
171
See, for example: IBM Corporation v. United States, 298 U.S. 131 (1936) (business machines and punch
cards); International Salt Co., Inc. v. United States, 332 U.S. 392 (1947) (salt machines and salt); McKie
[1955; 1959] (can-closing machines and tin cans); Advance Business Systems & Supply Co. v. SCM
Corp., 287 F.Supp. 143 (D.Md. 1968) (copiers and their supplies and service); In re Xerox Corp., 86 FTC
364 (1975) (copiers and their supplies and service).
172
Allen-Myland v. International Business Machines Corp., 33 F.3d 194 (3rd Cir. 1994) (recapturing old
parts in order to eliminate the competing secondhand market); Pacific Mailing Equipment Corp. v. Pitney
Bowes, Inc., 499 F.Supp. 108 (N.D. Cal. 1980) (destroying old machines in order to eliminate the
secondhand market). See also United Shoe Machinery (1953), at 344-345:
[The leasing practices] represent something more than the use of accessible resources, the
process of invention and innovation, and the employment of those techniques of employment,
financing, production, and distribution, which a competitive society must foster. They are
contracts, arrangements, and policies which, instead of encouraging competition based on pure
merit, further the dominance of a particular firm. In this sense, they are unnatural barriers;
they unnecessarily exclude actual and potential competition; they restrict a free market.
(emphasis added).
-84-
173
United Shoe (1953); McKie [1955, 1959]. As discussed, there are various ways by which a lease-only
practice may erect barriers to entry, e.g., long-term leases and forcing two-level entry through tying. The
latter may be facilitated by elimination of the secondhand market, which raises the costs of fringe firms
(R&D costs and replacement parts). In Van Dyk, the Court of Appeals for the Third Circuit rejected this
argument because the leases were short term. Research Corp. v. Xerox Corp., 631 F.2d 251 (3rd Cir.
1980). Bucovetsky & Chilton [1986] suggested that under a threat of entry, the durapolist improves his
position by selling rather than leasing. Their model premises that a lease-only policy is designed to
overcome time inconsistency and does not take into account other anticompetitive effects which deter
entry. In contrast to Bucovetsky & Chiltons argument, IBM did fight to maintain a high level of leasing,
although it was under a threat of entry. See: Telex Corp. v. International Business Machines Corp., 510
F.2d 894, 916-19 (10th Cir. 1975); Greyhound Computer v. International Business Machines, 559 F.2d
488, 498 (9th Cir. 1977):
There was ample evidence that IBM officials became concerned that the balance between
sales and rental had turned too heavily toward sales, and deliberately set about to reverse the
trend.
174
Bulow [1982, 1986]; Malueg & Solow [1987]; Waldman [1993; 1996]; Hendel & Lizzeri [1999].
-85-
this policy and reduced durability is unclear. USM, for instance, was known for frequent
introductions of improved models of its machines. Second, as discussed, both leasing
policies and planned obsolescence may be used to deter entry and to raise rivals costs.
Therefore, lease-only and planned obsolescence may go hand in hand, rather than
substitute each other. For instance, in many circumstances leasing practices lower the
costs of planned obsolescence because the innovative lessor can cheaply recycle the old
goods, or, alternatively, prevent his competitors from using old parts in order introduce
new, cheaper models.175
deadweight loss implicitly premises that the durapolists market power is a fixed sum.
Namely, it is assumed that the durapolist cannot leverage his market power to related
markets and it is assumed that there is no significance to entry barriers in terms of market
power.176 The durapolist can garner a higher reward if he employs a lease-only policy
with other business strategies, and therefore the deadweight loss argument has no sound
foundation.
Indeed, leasing even by the durapolist has important economic functions.
Nevertheless, these advantages do not justify lease-only policies and can be achieved by
leasing companies or with a mixed policy of selling-leasing by the durapolist.177 In
175
See subsection D.5. See also: Pacific Mailing Equipment (1980), supra note 172; Allen-Myland (1994),
supra note 172; Greyhound Computer v. International Business Machines, 559 F.2d 488, 498 n. 22 (9th
Cir. 1977):
Leasing was more advantageous to IBM than selling the same equipment [because it]
facilitated introduction of newly developed products, since lessees were not inhibited by a
large investment in either the new or the old machine.
176
For a presentation of the fixed-sum argument and its flaws, see Kaplow [1985]. For the relationship
between market power and barriers to entry, see: Baumol, Panzar & Willig [1988].
177
As noted, IBM has used various strategies to turn its mixed policy toward leasing. These efforts
apparently gained some success. Nevertheless, the existence of competing leasing companies and the
-86-
particular, generally there is no economic reason to allow the durapolist to offer only longterm leases, although the existence of switching costs and risk allocation considerations
may make these leases economically desirable. A long-term lease is plainly a product and,
therefore, if its marketing makes economic sense this product is expected to be on the
market.
The alternative seems to be undesirable, even assumed for the argument that a
lease-only policy presents no barriers to entry.
market for purchased mainframes suggest that although the durapolist cannot be perfectly monitored, his
market power is restricted when he is required to sell.
178
Wiley, Rasmusen & Ramseyer [1990], at 703: The horse we beat is not dead. But neither is it well
shod.
-87-
alive, but is also still kicking as in its youth although it is not the horse Professor Coase
sold us.
G.
Concluding Remarks
My concluding remarks start where I began and this may suggest that I have not
gone too far. Durapolists may have time inconsistencies which hinder them from charging
the static monopoly price. To overcome these problems, durapolists often engage in
various practices that bring them closer to the static monopoly price. This incremental
increase in the durapolists market power is a violation of 2 of the Sherman Act, in the
absence of proper justifications, such as encouraging R&D activities and efficient
management.
Most of the practices that I examined, were mentioned by Professor Coase in his
seven-page note and this may imply, again, that indeed I have not come a long way. My
points that are worth being stressed are as follows:
(a) Any regulation of durapolists practices must take into account its ex ante impacts on
potential durapolists. In particular, the impacts on investment decisions concerning
the creation of new markets or the development of existing ones.
(b) Under a durable-good monopoly, most-favored-nations clauses to a significant share
of the customers are used as devices to maintain market power.
(c) The social costs of durable-goods cartels are higher than is usually assumed, because
cartels have mechanisms in place to avoid time inconsistencies.
-88-
(d) Planned obsolescence is not always socially undesirable, but given verifiability, there
are circumstances of planned obsolescence with high social costs that justify antitrust
intervention.
(e) Tying may provide a device to overcome the durapolist problem.
(f)
-89-
REFERENCES
ADAMS, RUSSELL B., KING C. GILLETTE: THE MAN
AND
DEVICE (1978).
Aghion, Phillippe and Bolton, Patrick, Contracts as a Barrier to Entry, 77 AM. ECON.
REV. 388 (1987).
AREEDA, PHILIP E., ANTITRUST LAW: AN ANALYSIS
OF
ANTITRUST PRINCIPLES
AND
THEIR APPLICATIONS.
________________ and Turner, Donald F., Predatory Pricing and Related Practices
under Section 2 of the Sherman Act, 88 HARV. L. REV. 697 (1975).
ARTHUR, W. BRIAN, INCREASING RETURNS AND PATH DEPENDENCE IN THE ECONOMY
(1994).
Ausubel, Lawrence M. and Deneckere, Raymond, One is Almost Enough for
Monopoly, 18 RAND J. ECON 225 (1987).
Avinger, Robert L., Product Durability and Market Structure: Some Evidence, 29 J.
INDUSTRIAL ECON. 357 (1981).
Ayres, Ian, How Cartels Punish: A Structural Theory of Self-Enforcing Collusion, 87
COLUM. L. REV. 295 (1987).
Bagwell, Laurie S. and Bernheim, B. Douglas, Veblen effects in a theory of
conspicuous consumption, 86 AM. ECON. REV. 349 (1996).
Baker, Jonathan B., Fringe Firms and Incentives to Innovate, 63 ANTITRUST L. J. 621
(1995).
________________, Vertical Restraints with Horizontal Consequences: Competitive
Effects of Most-Favored-Customer Clauses, 64 ANTITRUST L. J. 517 (1996).
Barro, Robert J., Monopoly and Contrived Depreciation, 80 J. POL. ECON. 598 (1972).
-90-
AND
AND
THE
SUN?: PEOPLE,
AND
AND
ECONOMIC
APPRAISAL (1973).
Bresnahan, Timothy F., Post-Entry Competition in the Plain Paper Copier Market, 75
AM. ECON. REV. 15 (1985).
BRIGHT, ARTHUR A., THE ELECTRIC LAMP INDUSTRY TECHNOLOGICAL CHANGE
ECONOMIC DEVELOPMENT FROM 1800 TO 1947 (1949).
-91-
AND
OF
MARKET POWER
(1975).
Brodley, Joseph F. and Ma, Ching-To A., Contract Penalties, Monopolizing
Strategies, and Antitrust Policy, 45 STAN. L. REV. 1161 (1993).
Bucovetsky, Sam and Chilton, John, Concurrent Renting and Selling in a DurableGoods Monopoly under Threat of Entry, 17 RAND J. ECON. 261 (1986).
Bulow, Jeremy I., Durable-Goods Monopolists, 90 J. POL. ECON. 314 (1982).
______________, An Economic Theory of Planned Obsolescence, 101 Q. J. ECON.
729 (1986).
Bursten, Meyer L., The Economics of Tie-in Sales, 42 REV. ECON. & STAT. 68 (1960).
Butz, David A., Durable-Good Monopoly and Best-Price Provisions, 80 AM. ECON.
REV. 1062 (1990).
Carlton, Dennis W., The Rigidity of Prices, 76 AM. ECON. REV. 637 (1986).
_______________ and Gertner, Robert H., Market Power and Mergers in DurableGoods Industries, 32 J. L. & ECON. S203 (1989).
Celnicker, Arnold, A Competitive Analysis of Most Favored Nations Clauses in
Contracts between Health Care Providers and Insurers, 69 N.C. L. Rev. 863 (1991).
Coase, Ronald H., Durability and Monopoly, 15 J. L. & ECON. 143 (1972).
Cooper, Thomas E., Most-Favored-Customer Pricing and Tacit Collusion, 17 RAND J.
ECON. 377 (1986).
________________ and Fries Timothy L., The Most-Favored-Nation Pricing Policy
and Negotiated Prices, 9 INTL J. INDUS. ORG. 209 (1991).
Craswell, Richard, Tying Requirements in Competitive Markets: The Consumer
Protection Issues, 62 B. U. L. REV. 661 (1982).
David, Paul A., Clio and the Economics of QWERTY, 75 AM. ECON. REV. 332 (1985).
-92-
DeGraba, Patrick, No Lease is Short Enough to Solve the Time Inconsistency Problem,
42 J. INDUS. ECON. 361 (1994).
DELAMARTER, RICHARD T., BIG BLUE: IBMS USE AND ABUSE OF POWER (1986).
Denicolo, Vincenzo and Garella, Paolo G., Rationing in a Durable Goods Monopoly,
Dennis, Anthony J., Potential Anticompetitive Effects of Most Favored Nation
Contract Clauses in Managed Care and Health Insurance Contracts, 4 ANNALS
HEALTH L. 71 (1995a).
_________________, Most Favored Nation Contract Clauses under the Antitrust
Laws, 20 U. DAYTON L. REV. 821 (1995b).
Director, Aaron and Levi, Edward H., Law and the Future: Trade Regulation, 51 NW.
U. L. REV. 281 (1956).
DIXIT, AVINASH K.
AND
ON
WAR MOBILIZATION
OF THE
SENATE COMM.
ON
ND
AND
-93-
-94-
Hay, George A., Facilitating Practices: The Ethyl Case (1984), in THE ANTITRUST
REVOLUTION: ECONOMICS, COMPETITION,
AND
OF THE
ANTITRUST
(1994).
V.
-95-
Klein, Benjamin, Market Power in Antitrust: Economic Analysis after Kodak, 3 SUP.
CT. REV. 43 (1993).
Klemperer, Paul, Markets with Consumer Switching Costs, 102 Q. J. ECON. 375
(1987).
Krattenmaker, Thomas G. and Salop, Steven C., Anticompetitive Exclusion: Raising
Rivals Costs to Achieve Power over Price, 96 YALE L. J. 209 (1986).
Kreps, David M. and Wilson, Robert B., Reputation and Imperfect Information, 27 J.
ECON. THEORY 253 (1982).
Kwoka, John E., The Sales and Competitive Effects of Styling and Advertising in the
U.S. Auto Industry, 75 REV. ECON. STAT. 649 (1993).
Lazear, Edward P., Retail Pricing and Clearance Sales, 76 AM. ECON. REV. 14 (1986).
Lee, In Ho and Lee Jonghawa, A Theory of Planned Obsolescence, 46 J. INDUS. ECON.
383 (1998).
Leibenstein, Harvey, Bandwagon, Snob, and Veblen Effects in the Theory of
Consumers Demand, 64 Q. J. ECON. 183 (1950).
LENZEN, GODEHARD, THE
TRADE (F.
Liebowitz, S. J., Durability, Market Structure, and New-Used Goods Models, 72 AM.
ECON. REV. 816 (1982).
_____________ and Margolis, Stephen E., Network Effects and Externalities, in THE
NEW PALGRAVE DICTIONARY
OF
ECONOMICS
AND THE
-96-
386 (John E. Kwoka and Lawrence J. White eds., 3rd ed., 1999).
Malueg, David A. and Solow, John L., On Requiring the Durable Goods Monopolist
to Sell, 25 ECON. LETTERS 283 (1987).
Marvel, Howard P. and Peck, James, Demand uncertainty and returns policies, 36
INTL ECON. REV. 691 (1995)
Masten, Scott E. and Snyder, Edward A., United States versus United Shoe Machinery
Corporation: On the Merits, 36 J. L. & ECON. 33 (1993).
McAffe, R. Preston and Schwartz, Marius, Opportunism in Multilateral Vertical
Contracting: Nondiscrimination, Exclusivity, and Uniformity, 84 AM. ECON. REV. 210
(1994).
McGee, John S., Economics of Size in Auto Body Manufacture, 16 J. L. & ECON. 239
(1973).
_____________, Compound Pricing, 25 ECON. INQ. 315 (1987).
MCKIBBEN, GORDON, CUTTING EDGE: GILLETTES JOURNEY
TO
GLOBAL LEADERSHIP
(1998).
McKie, James W., The Decline of Monopoly in the Metal Container Industry, 45 AM.
ECON. REV. 499 (1955).
______________, TIN CANS
AND
OF
COMPETITION
IN
TWO
-97-
OBrien, Daniel P. and Shaffer Greg, Vertical Control with Bilateral Contracts, 23
RAND J. ECON. 299 (1992).
Ordover, Janusz A. and Willig, Robert D., An Economic Definition of Predation:
Pricing and Product Innovation, 91 YALE L. J. 8 (1981).
_______________, Sykes, Alan O. and Willig, Robert D., Predatory System Rivalry,
83 COLUM. L. REV. 1150 (1983).
_______________,________________and________________,
Anticompetitive
Behavior
by
Dominant
AND
Firms
Toward
Nonprice
the
REGULATION: ESSAYS
Producers
IN
MEMORY
of
OF
OF INDUSTRIAL
OF
(Sep. 1993).
_____________ and _______, Manufacturers Returns Policies and Retail
Competition, 16 MARKETING SCI. 81 (1997).
Pashigian, B. Peter, Bowen, Brian and Gould Eric, Fashion Styling and the WithinSeason Decline in Automobile Prices, 38 J. L. & ECON. 281 (1995).
Pesendorfer, Wolfgang, Design Innovation and Fashion Cycles, 85 AM. ECON. REV.
771 (1995).
Petty, Ross D., Antitrust and Innovation: Are Product Modifications Ever Predatory?,
22 SUFFOLK U. L. REV. 997 (1988).
-98-
AND
-99-
OF
AND
IN
ACTION: CASE
-100-
OF
IN
OF
RESEARCH,
Optimum
Durability,
Second
Hand
Markets,
and
Planned
-101-
Wiley John S., Rasmusen Eric B., and Ramseyer J. Mark, The Leasing Monopolist, 37
UCLA L. REV. 693 (1990).
WILLIAMSON, OLIVER E., MARKETS
AND
HIERARCHIES: ANALYSIS
AND
ANTITRUST
IMPLICATIONS (1975).
___________________, THE ECONOMICS INSTITUTIONS OF CAPITALISM (1985).
-102-